summary

This project aims to carry out systematic pathogenic mechanism analysis and new diagnosis and treatment strategies for common mental disorders such as major depressive disorder, bipolar disorder and schizophrenia. Based on the DIKWP model (Data-Information-Knowledge-Wisdom-Purpose) and artificial consciousness theory proposed by Professor Duan Yucong, the project integrates neurobiology, multi-omics, big data analysis and social psychology methods to explore the multi-scale mechanisms of mental disorders from the cellular and molecular level to the information processing and consciousness level. The main research contents include: drawing disease-specific cell maps and neural circuit structures in brain regions, and deducing the mechanism of lesion regional information processing defects and consciousness disturbance based on the DIKWP model; Elucidate the abnormal pathogenic pathways of mental disorders, and use the evidence of effective clinical drugs to reversely verify the relationship between the etiological mechanism of the disease and the hypothesis of "information control failure" in artificial consciousness. To explore the objective biomarkers of early diagnosis of mental disorders, and to establish a prediction index system based on the three-level pathway of cognition-information-structure. To study the role of social stress and psychological intervention in the occurrence and development of diseases, and to explore the linkage mechanism between social stress and the "environmental mapping layer" in the evolution of artificial consciousness. and exploring a new comprehensive treatment model that combines pharmacotherapy with psychosocial interventions. The project will be designed in accordance with the specifications of national scientific research projects, and adopt multidisciplinary and interdisciplinary technical routes and innovative methods. The expected outputs include revealing the key pathogenic circuits and abnormal patterns of information processing in mental disorders, proposing new etiological hypotheses and diagnostic indicators from the perspective of artificial consciousness, and developing new models and intervention strategies for comprehensive treatment. The results of this study will help deepen the understanding of the nature of mental disorders, promote more effective early diagnosis and individualized treatment plans, and provide a scientific basis for improving the prognosis and quality of life of patients.

Basis for the project

1. Current status and research challenges of mental disorders

Depression, bipolar disorder, and schizophrenia are major mental disorders with high prevalence and disability worldwide. Depression alone is estimated by the World Health Organization to affect about 280 million people and is one of the leading causes of disability worldwide, bipolar disorder affected about 40 million people in 2019, and schizophrenia affects about 24 million people. These diseases usually occur in young adults and have the characteristics of prolonged recurrence, which brings a heavy burden to the patient's family and society. Although there are some pharmacological and psychotherapeutic approaches for these disorders, the efficacy is limited, and many patients struggle to achieve complete remission or recovery. As an example, about one-third of patients with depression do not respond significantly to initial antidepressant therapy, and the recurrence rate of treatment-resistant depression is as high as 75%. Patients with bipolar disorder have a poor long-term prognosis and significant functional impairment; The life expectancy of patients with schizophrenia is shortened by 10-20 years compared with ordinary people, which seriously affects the quality of life. Obviously, the existing treatment methods cannot fully solve the global public health problem of mental disorders, and it is urgent to deepen the understanding of disease mechanisms through basic research, so as to develop new diagnosis and treatment strategies.

For a long time, the pathogenic mechanism of mental disorders has been thought to involve the interaction of complex biological, psychological, and social factors. At present, the academic community still has insufficient understanding of these diseases, and there are many key scientific questions to be answered. For example: Do different mental disorders have a common neurobiological basis? Why do these disorders cause abnormalities in cognitive function and mental status? What specific areas of the brain, cell types, and circuits are affected in disease? How does environmental stress induce or exacerbate the disease biologically? Research on these issues is not only of great scientific importance, but also a basis for the development of new diagnostic and therapeutic technologies. However, due to the high complexity of the human brain and the high diversity of psychiatric disorders, it has been difficult to fully unravel the cross-level mechanisms from cellular molecules to system behavior. In recent years, with the rapid development of neuroscience and omics technology, we have the opportunity to deeply characterize the disease brain at the single-cell and circuit level, bringing revolutionary breakthroughs to the study of mental disorders.

2. The development of multi-scale brain science and technology provides new opportunities for mechanistic research

Cell mapping and circuit elucidation in brain regions: With the help of single-cell transcriptome sequencing, large-scale brain tissue imaging, and connectome mapping, researchers have begun to construct cell types and circuit connectivity maps of the human brain. This lays the foundation for understanding the microscopic changes in the brain in psychiatric disorders. Recently, a large study analyzed more than a million human brain cells, mapped the distribution of 107 cell subtypes in 42 brain regions, and associated specific cell types with diseases such as schizophrenia, bipolar disorder, and depression. For example, Nature Neuroscience published results that show that risk genes for schizophrenia, bipolar disorder, and depression are enriched in specific neuronal populations, such as somatic inhibitory interneurons (SST-positive cells), excitatory neurons in the posterior cingulate cortex, and medium-sized spinous projection neurons in the amygdala. This suggests that there may be different "clusters of diseased cells" and "abnormal circuit nodes" in the brain in different diseases. At the same time, at the brain network level, functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) studies have revealed that patients with mental disorders often have abnormal brain functional network connectivity. For example, in patients with schizophrenia, the Default Mode Network (DMN) has both weakened and increased connections to other brain networks, and this disconnection may be an important cause of symptoms such as hallucinations, delusions, and confusion. Patients with depression often present with hyperactivity of DMN (associated with introverted contemplation) and a low cognitive control network, resulting in negative emotional rumination and attentional bias towards negative information. These findings suggest that advanced multi-scale brain science and technology can be used to deeply characterize cell-circuit-network abnormalities related to mental illness, which lays a research foundation for this project.

Abnormal information processing and alteration of consciousness: Mental disorders are not only anatomical and biochemical changes in the brain, but also manifest as abnormalities in information processing function and state of consciousness. For example, patients often have cognitive symptoms such as inability to concentrate, oversensitivity or sluggishness, impaired thinking and association, and impaired control of volitional behavior. A growing body of research has identified these as defects in the brain's information processing process. Some scholars have proposed that schizophrenia can be regarded as a "consciousness disorder", partly due to the disconnection between unconscious information processing and conscious cognition, which makes it difficult for patients to filter sensory input and automatically process daily information, and has to make up for it with conscious efforts, resulting in abnormal experiences. Experimental evidence shows that patients with schizophrenia have early sensory gating disorder and attentional filtering deficits, and are unable to inhibit irrelevant stimuli, resulting in perceptual overload and cognitive fragmentation. Similarly, patients with major depressive disorder exhibit an attentional bias and cognitive rigidity towards negative emotional information, which can be seen as the brain's inability to properly screen and control the flow of information—positive information is not effectively transmitted to high-level cognition, and negative information is over-amplified. In the manic episode stage of bipolar disorder, it is common to have rushing thoughts, excessive motivation, and weakened impulse control, which can be understood as an imbalance in the regulation of information and motivation at the higher control level. These phenomena coincide with the "information control failure" hypothesis: that is, the occurrence of mental disorders is likely to stem from the dysfunction of the brain's key information processing/control circuits, resulting in deviations or loss of control of the cognitive system's filtering, integration, and interpretation of internal and external information. This project intends to use this as a starting point to characterize the manifestation of this information flow disorder in diseases with the help of the consciousness model in artificial intelligence, and to answer the mechanism why the failure of information control can lead to consciousness and behavior abnormalities.

3. A new theoretical perspective introduced by the DIKWP artificial consciousness model

Traditional research mostly focuses on biology or psychology, and the understanding of mental disorders is relatively fragmented. This project innovatively introduces Professor Duan Yucong's DIKWP artificial consciousness theoretical framework to provide a holistic cognitive perspective that integrates data-information-knowledge-wisdom-purpose for the study of mental disorders. The DIKWP model is based on the classic DIKW (pyramid) model, which adds a "Purpose" layer, which is composed of five levels: data→ information→ knowledge→ Wisdom → Purpose. This model emphasizes the guiding and constraining role of the highest level of purpose on the cognitive processing of the lower level, and realizes the dynamic regulation and self-renewal of the cognitive process through two-way feedback between the levels. The research of Professor Duan Yucong's team shows that the DIKWP model provides an interpretable and controllable cognitive architecture for artificial intelligence systems, and has made breakthroughs in artificial consciousness evaluation by embedding the "Purpose" layer to achieve the alignment of machine decision-making to the goal. More importantly, the model attempts to simulate the process of human consciousness formation, so that AI has human-like self-monitoring and reflection capabilities. This provides a new way to re-examine mental disorders from the perspective of information science and systems theory: perhaps mental illness can be seen as a dysfunctional manifestation of the DIKWP architecture in the biological brain- That is, there are bugs (defects) in information processing at different levels or an imbalance in feedback between levels, resulting in exceptions at the Wisdom decision-making and Purpose level. This project will use the DIKWP model as a theoretical tool to systematically analyze the possible problems in the brain of patients with mental disorders in data acquisition, information integration, knowledge refinement, wisdom decision-making and purpose control. For example, is the data layer of the sensory input already noise-filtering? Is the processing of perception signals distorted at the information level? Does the knowledge layer (the brain's intrinsic model) form false causal inferences (such as delusions)? Are the decision-making functions of the Wisdom layer (e.g., situational judgment and value evaluation) impaired? And is the highest level of purpose (motivation/self-will) unable to effectively guide the lower levels, causing the patient to lose control of his thoughts and actions? These are all scientific questions that need to be explored urgently.

In addition, the DIKWP artificial consciousness framework proposes a "subconscious + conscious dual cycle" structure, that is, in addition to the basic cognitive process, a metacognitive cycle is also introduced to achieve self-monitoring and regulation. This is similar to the human brain's conscious and subconscious feeding mechanism. Duan Yucong's team also proposed the "BUG" theory of consciousness, arguing that a moderate amount of randomness and imperfection should be allowed in artificial consciousness to simulate the creativity and robustness of human thinking. By analogy to mental illness, we can reflect: Is it possible that the "bugs" of the human brain, the tiny errors in the processing of information caused by certain genes or developments, may be accumulated and amplified into psychopathological phenomena? Some hypotheses suggest that people with schizophrenia have an elevated level of intrinsic noise in brain signals, leading to erroneous correlations being processed by the brain as really important information (i.e., abnormal "salient conferencing"). The "deviant significance" hypothesis proposed by Kapur et al. refers to the fact that the striatal dopamine dysfunction causes irrelevant stimuli to be given excessive meaning and triggers delusional hallucinations. This view is also essentially a manifestation of the failure of information control. In depression, the brain may have insufficient saliency to positive stimuli, and be overly sensitive to negative implicit information, resulting in negative cognitive paranoia. Different phases of bipolar disorder may oscillate between the two extremes of the saliency evaluation mechanism. This project will combine the DIKWP model and artificial consciousness theory to systematically demonstrate and verify the above conjectures. Through this interdisciplinary theoretical fusion, we hope to provide a full-link mechanistic explanatory framework for mental disorders**, connecting molecular, cellular, and circuit-level pathologies with information processing, cognitive function, and subjective conscious experience.

4. The new need for accurate diagnosis and comprehensive intervention

Timely and accurate early diagnosis and scientific comprehensive treatment are the keys to improving the prognosis of mental disorders. In reality, however, there are significant deficiencies in both of these aspects. At present, clinical diagnosis is mainly based on subjective assessment of symptomatology, and there is a lack of objective biomarkers. Underlying changes in the brain may have occurred over a long period of time before the patient develops noticeable symptoms. Early markers that predict a high risk of disease can be identified to significantly advance the window of intervention. Studies have explored the combination of multimodal biomarkers, such as cognitive testing + brain imaging + genetic information, to improve the accuracy of predicting the occurrence of mental illness. A study in people at high risk of psychosis showed that machine learning using multimodal brain signal fusions such as structural MRI, functional connectivity, and diffusion tensor imaging could predict whether an individual would transition from high risk to schizophrenia episodes more accurately than a single modality (multimodal model AUC=0.73, significantly better than any unimodal AUC~0.66). This shows that the three-level path prediction method of "cognition-information-structure" is feasible and superior. Focusing on this idea, this project plans to systematically explore the early warning indicators of mental disorders: at the cognitive level, quantify the decline of small memory, attention, social cognition and other functions; At the information level, neuroelectrophysiological signals such as EEG/P300 and objective data such as eye movement and physiological sensing were used to capture abnormal patterns of information processing. At the structural level, brain imaging, biomarkers (e.g., inflammatory factors, hormone levels), and genetic/epigenetic markers are incorporated. Through machine learning and semantic network analysis, these different levels of information are fused to build predictive models and evaluate their sensitivity and specificity. We are particularly interested in the association between cognitive-information-structure pathways: for example, does the brain image of an individual with mild cognitive deficit and abnormal levels of stress hormone already show altered connectivity to specific networks? Can these changes predict future disease progression? Through in-depth research, we hope to find an objective and actionable combination of early diagnostic markers** to inform clinical screening and intervention.

On the other hand, psychosocial factors play an important role in the occurrence and prognosis of mental disorders, which cannot be ignored. Studies have shown that experiences such as childhood trauma and chronic social stress can have a profound impact on the brain through epigenetics and other pathways, making individuals more susceptible to depression, anxiety, and psychotic disorders in adulthood. Longitudinal cohort studies have shown that people who have suffered multiple forms of childhood abuse have a significantly higher risk of depression in adulthood, and tend to have an earlier onset, a more chronic course of the disease, and a worse response to treatment. Social isolation, discrimination, and the stress of urbanizing life are associated with an increased risk of developing schizophrenia: the risk of developing schizophrenia is about 2.7 times higher in first-generation immigrants than in the local population, and up to 4.5 times higher in second-generation migrants, which is thought to be related to the social exclusion and stress experienced by immigrants. It can be seen that environmental stress profoundly affects mental health through various mechanisms such as chronic activation of the hypothalamic-pituitary-adrenal (HPA) axis, induction of neuroinflammation, and shaping cognitive beliefs. This project will delve into the mechanisms by which social stress affects the brain and consciousness. On the one hand, at the animal level, we will use the chronic stress model and the mouse social frustration model to simulate the effects of environmental stress on neural circuitry and gene expression in the brain, and observe whether there are behavioral and neurobiological changes similar to human mental disorders. On the other hand, in clinical studies, we will combine epidemiological and neuroimaging data to analyze whether there are specific changes in brain structure and function, such as a decrease in hippocampal volume (a classic effect of chronic stress) and the amygdala, in people who have experienced high-pressure life events- Dysfunction of functional connectivity of the prefrontal lobes, etc. In addition, we plan to explore the mechanisms of action of psychological interventions in reversing or buffering these alterations. Evidence has shown that cognitive behavioral therapy (CBT) and mindfulness-based stress reduction training can alter the pattern of brain functional activity, such as reducing the overconnection of the default network in patients with depression and enhancing the activity of the cognitive control network, thereby reducing symptoms. Family intervention and social skills training can reduce the relapse rate in schizophrenia, and their effect is partly attributed to improving patients' ability to cope with environmental stress and cognitive restructuring.

We will introduce the concept of "environmental mapping layer" in artificial consciousness theory**, that is, the consciousness system needs to effectively model and map the external environmental information in order to produce adaptive cognitive responses. In AI, this means that AI needs to map perceived environmental states to internal semantic representations so that it can adjust its actions accordingly. In the same way, the human brain continues to construct internal representations of the social environment in which it lives, and responds cognitively and emotionally based on these representations. If the environment mapping is skewed, such as mistaking a safe environment for a threatened environment, it can lead to a persistent stress response and pathological psychology. This project hypothesizes that long-term social stress will cause the "inner world model" to deviate from reality by interfering with the brain's environmental mapping representation, thereby triggering symptoms such as anxiety and paranoia. Effective psychological interventions are equivalent to recalibrating the brain's environmental mapping, so that patients have a more accurate and adaptive conscious response to external stimuli. This hypothesis will be tested in our study by longitudinal follow-up and intervention experiments. The ultimate goal of the project is to combine drug treatment with psychological intervention**** and implement integrated treatment for bio-information-social factors, in order to achieve better outcomes and prognosis than treatment alone.

In summary, the scientific basis for the establishment of this project is that the occurrence of mental disorders is the result of the joint action of multi-level pathological changesIt needs to be comprehensively analyzed by new theoretical frameworks and research methods. The DIKWP artificial consciousness model provides us with a theoretical perspective that runs through data to purpose, and can link neurobiological mechanisms to cognitive function and consciousness phenomena. Modern multi-omics and brain imaging techniques have made it possible to conduct cross-scale research. On this basis, we will hopefully answer a series of major scientific questions in the field of mental disorders and bring breakthrough ideas for improving diagnosis and treatment. Therefore, the development of this project has important scientific significance and potential social benefits.

Research Objectives:

The overall goal of this project is to elucidate the key links related to information processing and awareness in the pathogenesis of common mental disorders such as major depressive disorder, bipolar disorder, and schizophrenia, and to explore innovative diagnostic markers and treatment strategies accordingly. Specific objectives include:

1. Characterizing disease-specific cell maps and circuit structure abnormalities in brain regions: Using single-cell sequencing and brain imaging techniques, we can map the distribution of cell types and neural circuit connections in brain regions associated with depression, bipolar, and schizophrenia, and identify disease-specific "foci" cell populations and abnormal brain networks. At the same time, the DIKWP model was used to analyze the information processing function of these brain regions in normal cognition, and the mechanism of information flow obstruction or distortion in disease states was inferred.

2. To analyze the abnormal pathogenic mechanism of mental disorders and verify the hypothesis of "information control failure": to explore how the abnormalities of these pathways lead to cognitive and conscious dysfunction by targeting the typical biological pathways in three diseases (such as the monoamine and neurotrophic mechanisms of depression, the GSK-3 signaling pathway of bipolar disorder, and the dopamine-glutamate imbalance of schizophrenia), combined with the targets of clinically effective drugs (such as SSRIs, lithium salts, and antipsychotics). Using the evidence of drug reversal abnormalities, the role of the failure of key information control links in the brain in the etiology of diseases was verified. For example, do SSRIs increase 5-HT levels to alleviate depression by restoring normal processing of emotional information? Does lithium salt enhance cognitive regulation by inhibiting GSK-3β to promote neuroplasticity? Whether antipsychotic blockade of dopamine D2 receptor reduction delusions confirms correction of excessive "salient" markers.

3. Exploring biomarkers for early diagnosis of mental disorders: Establish a three-level index system of "cognition-information-structure" to screen the combination of objective markers that can predict the high-risk state or early prognosis of the disease. Focus on: mild patterns of cognitive impairment, changes in brain network functional connectivity (e.g., dysregulation of the default network and task-positive network), minor abnormalities in brain structure (e.g., subtle changes in cortical thickness or hippocampal volume), and levels of key proteins or metabolites (e.g., stress hormones, inflammatory factors). Apply multimodal data fusion and machine learning to develop predictive models and evaluate their usefulness. The goal is to find sensitive and reliable indicators for early warning and graded diagnosis of disease.

4. Elucidate the mechanism of social stress and psychological interventions: Through a combination of animal and clinical studies, we will reveal how chronic stress induces or aggravates mental disorders through neuroendocrine, immune, and neural circuit shaping pathways. Based on the theory of "environmental mapping layer" of artificial consciousness, this paper analyzes the impact of long-term stress on the internal environment model of the brain, and how effective psychological intervention can reshape this model to make it more adaptable to reality. Specific objectives include: verifying epigenetic alterations and changes in brain structure/function caused by stressors such as childhood adversity; To observe the modification of brain functional connectivity and cognitive patterns by psychological interventions (e.g., CBT, family therapy), and the correspondence between such changes and symptom improvement; To explore the protective mechanism of social support for disease prognosis.

5. Exploring innovative treatment models combining drugs and psychological interventions: designing and evaluating bio-psychological integrated treatment programs based on mechanistic studies。 To assess the effects of antidepressants combined with cognitive behavioural therapy to improve cognitive biases and prevent relapse for depression; For bipolar disorder, the effect of mood stabilizers combined with family psychoeducation in prolonging remission was evaluated; To assess the effects of antipsychotics combined with social functioning training/vocational rehabilitation on improving functional outcomes for schizophrenia. This paper summarizes the synergistic mechanism and applicable population of each combination of therapies, and forms a comprehensive intervention model that can be promoted to provide a basis for clinical practice. The goal is to demonstrate that combination therapies are more effective than monotherapy in improving efficacy and adherence, and to elucidate the synergistic mechanisms behind them (e.g., the dual role of pharmacological improvement of biological abnormalities + psychological intervention to correct cognitive distortion).

By achieving the above goals, this project will comprehensively reveal the cross-scale mechanisms of mental disorders from microscopic cells to macroscopic behaviors, validate new etiological and treatment hypotheses, and lay a scientific foundation for the future development of objective diagnostic tools and comprehensive treatment options.

Research content and technical route

Focusing on the research objectives, the project sets up five major research contents (topics), and formulates corresponding technical routes and research plans for each part. Each part not only carries out in-depth research relatively independently, but also closely connects through the theoretical framework of DIKWP to form an overall interpretation of the pathogenic mechanism of mental disorders. The specific plans and technical roadmaps for each research content are described in the following order.

Research content 1: Characterization of cell maps and circuit structures in disease-specific brain regions

Main scientific question: What regions and cell types in the brain of people with mental disorders have specific changes? How do these changes affect the brain's information processing processes and disrupt the state of consciousness?

Research Protocol: This content will use single-cell omics and brain imaging techniques to provide an in-depth analysis of the brain of depression, bipolar disorder, and schizophrenia at the cell-circuit level. The subjects of the study included autopsy brain tissue of patients, living brain tissue samples obtained by craniotomy (e.g., those with depression in epilepsy surgical resection specimens of drug-resistant patients), and multimodal brain imaging data of living patients. The specific steps are as follows:

1. Single-cell transcriptome sequencing (scRNA-seq) mapping: Select key brain regions that are closely related to each disease to sample and isolate cells (e.g., prefrontal cortex and hippocampus for depression, nucleus accumbens and amygdala for bipolar disorder, hippocampus, anterior cingulate cortex, thalamus for schizophrenia, etc.). RNA sequencing of individual cells yields gene expression profiles from hundreds of thousands or even millions of cells. Bioinformatics clustering classified different cell types and subtypes, and compared the differences in the number of cells and gene expression between patients and healthy controls. The expected outcome is to identify disease-related cellular abnormalities, such as a significant decrease in the number or abnormal gene expression of certain inhibitory interneurons, and an increase in glial activation. Combined with existing studies, we pay special attention to the previously mentioned cells that have a role in a variety of diseases, such as SST-positive inhibitory interneurons, excitatory projection neurons, and dopaminergic neurons. If sequencing results also show abnormalities in these cells (e.g., clusters of excitatory neurons in the amygdala of bipolar and schizophrenia patients rich in disease-risk genes), their functional significance will be investigated in depth.

2. Spatial transcriptome and immunohistochemistry validation: In order to understand the spatial distribution of cell abnormalities and the corresponding anatomical circuits, spatial transcriptome sequencing or high-throughput in situ hybridization can be used to determine the gene expression distribution while maintaining the tissue structure in the above-mentioned brain regions, so as to map the anatomical map of cell abnormalities. At the same time, multiplex fluorescence immunohistochemistry/in situ hybridization was applied to validate the localization of key discovered cell types. For example, if scRNA-seq shows a decrease in cingulate gyrus lamellar SST cells in patients with schizophrenia, we will confirm the density change of SST+ neurons at that level by counting SST+ neurons by immunostaining. Combined with high-resolution microscopy and tissue clearing technology, the local loop structure can be reconstructed in 3D to observe potential differences in synaptic connections between disease and control (e.g., dendritic spine density, excitability/inhibition synaptic ratio, etc.).

3. Whole brain neuroimaging connectome analysis: For brain regions and circuits that are found to be abnormal at the single-cell level, we will verify their macroscopic impact in living patients using imaging methods such as MRI/DTI/fMRI. These include:

• Structural MRI and DTI: Measure gray matter volume, cortical thickness, and integrity of associated white matter fiber bundles in target brain regions. If single-cell results suggest neuronal impairment of the prefrontal-hippocampal circuit in patients with depression, we expect that a thinning of the prefrontal cortex, a decrease in hippocampal volume, and a decrease in the white matter tract anisotropy of the prefrontal-hippocampal line may be observed on structural images.

• Functional MRI (resting and tasking): focuses on the strength of the connections of the functional network in which the target brain region is located. For example, the anterior cingulate - default network, parietal lobe - note if the network is too strong or too weak in the patient. Based on our preliminary hypothesis, abnormal DMN sub-network connectivity in patients with schizophrenia may be associated with hallucinatory delusions; In patients with depression, DMN is inversely correlated with cognitive control network, weakened, and related to contemplation. These imaging indicators were correlated with the patient's cognitive phenotype (neuropsychological test results) and symptom severity to establish a cell-circuit-function link.

• EEG/Magnetic Encephalography (MEG): Captures the dynamics of information processing in abnormal brain regions in the temporal dimension. For example, when sensory gating is assessed with EEG ERP components (P300, MMN, etc.), a decrease in P300 amplitude in people at high risk of schizophrenia is considered to be one of the signals before the onset of positive symptoms. This project will test whether these biomarkers are associated with the specific brain region/cell abnormalities we find, such as whether the alteration of P300 is due to the abnormal synchronization of neuronal activity in a specific brain region.

 Inferring information processing deficiencies in combination with the DIKWP model: Based on the acquisition of disease-related cell and circuit abnormalities, we will use the DIKWP model to explain these brain changes at the functional level. This is done by mapping brain regions and their connections to different levels of DIKWP functions. For example, the sensory cortex corresponds to the "data" layer and is responsible for raw signal acquisition; The hippocampus and associated thalamic nuclei correspond to the "information" layer, which is responsible for memory and contextual information integration. The prefrontal cortex can correspond to the "Wisdom" layer, which is responsible for rule extraction, decision-making, and value judgment. The parietal-prefrontal attention network can be regarded as an information selection mechanism under the regulation of the "Wisdom" layer. The self-correlation processing of the default network and the motivation evaluation of the limbic system are related to the "Purpose" layer. With the help of this correspondence, we can deduce the problems that arise at each level of the disease state: for example, if there is a filtration disorder in the information layer (thalamic-sensory cortex circuit) in schizophrenia, then a large amount of irrelevant data will pour into the knowledge layer, resulting in a bias in the model construction (understanding of reality) in the knowledge layer, and it will be difficult for the wisdom layer to make correct judgments based on this, and finally the high-level goals (such as self-concept) in the purpose layer may also be confused. This deduction will be combined with computational modeling: we plan to build a simplified DIKWP computational model to simulate the flow of information in both normal and disease states. By artificially introducing specific circuit defects (such as reducing inhibition feedback, increasing noisy signals, etc.) into the model, and observing the changes in the model's output (such as knowledge reasoning error rate, purpose target bias), we can verify whether a certain type of brain circuit abnormality is enough to cause similar psychiatric symptoms. This computational simulation can be implemented either in a computer (with the DIKWP concept incorporated into the artificial neural network architecture) or with the help of a dynamic causal model (DCM) to verify the change in the direction of information flow on the fMRI data.

Technology Roadmap: The technical path of research content 1 integrates four stages: "tissue sample processing > single-cell/spatial omics, > image validation, > theoretical model analysis". We will first obtain brain tissue samples (ethically, obtained from brain banks and surgical procedures), perform single-cell sequencing and spatial transcriptome experiments, and find cellular and molecular abnormalities; Histology was then used to verify the authenticity and localization of the cell abnormality. In parallel, MRI/EEG examinations were performed on supporting clinical subjects, and imaging and electrophysiological features were extracted to corroborate with histological results. Finally, in the stage of theoretical modeling, the empirical findings are integrated into the cognitive model of the DIKWP framework for deductive simulation, and the theoretical explanation of the disease information processing disorder is proposed. Such multiple layers of validation ensure consistency from micro to macro and map biological abnormalities to cognitive deficits, providing a basis for subsequent research.

Research content 2: Analysis of abnormal pathogenic mechanism and verification of the hypothesis of "information control failure".

Main scientific question: What are the core pathophysiology of mental disorders? How can abnormalities in molecular pathways be correlated with impairment of cognitive/conscious function? Can the mechanism of action of clinically effective drugs reflect the causal mechanism of the disease, so as to verify the hypothesis of "information control failure"?

Research Protocol: This section will focus on the known important biological mechanisms of depression, bipolar disorder, and schizophrenia, and delve into how their abnormalities affect neurological information processing. At the same time, the hypothesis was reversely tested through drug intervention experiments: if the symptoms can be alleviated by correcting the abnormality of a certain link, it means that the abnormality of the link may be the key to pathogenesis. The specific plan is as follows:

1. Mechanism analysis of depression: Focusing on the two classic hypotheses of monoamine neurotransmitter imbalance and neurotrophic disorder. The levels of serotonin (5-HT), norepinephrine, dopamine and their metabolites were detected in peripheral blood and cerebrospinal fluid samples, and the possession of 5-HT transporters and receptors in the brain of depressed patients was evaluated by PET imaging, and the degree of monoamine signal disorder was confirmed. Simultaneous analysis of peripheral inflammatory factors (e.g., IL-6, TNF-α) and brain-derived neurotrophic factor (BDNF) levels was performed to verify whether depression was accompanied by chronic inflammation and decreased BDNF. Animal model experiments (a mouse model of depression induced by chronic unpredictable stress) were then conducted to observe how stress led to changes in monoamines and BDNF, as well as consequent behavioral abnormalities and neuroplastic alterations (e.g., reduction in dendritic spines in hippocampal neurons). Hypothesis validation of information control failure: an antidepressant intervention model animal and patient to test its effect on information processing and neuroplasticity. For example, can a selective 5-HT reuptake inhibitor (SSRI) restore dendritic spine density and cognitive flexibility in hippocampal neurons in model mice? Studies have shown that chronic stress leads to a decrease in the synaptic ridge and impaired neuroplasticity, and rapid antidepressants such as ketamine can reverse synaptic function and increase BDNF levels within 24 hours, thereby rapidly improving symptoms. We will replicate this phenomenon in the model and further explore: Does ketamine's increased BDNF-TrkB signaling and induced synaptic enhancement correspond to the recovery of processing capacity in the knowledge/wisdom layer (e.g., improved performance of animals on cognitive tasks) in the DIKWP model? If the answer is positive, the core cause of depression is a decline in the plasticity of the brain's information processing network, and a sluggish control of information, which is corrected by drugs that increase plasticity (increase the efficiency of "data-information" conversion).

2. Mechanism analysis of bipolar disorder: focusing on intracellular signal transduction and neuroplastic rhythm abnormalities. Bipolar disorder is unique in that the patient's neurobiological status fluctuates between the depressive and manic poles, suggesting a possible failure of the "timing control" mechanism that regulates homeostasis. We will investigate the role of GSK-3β signaling pathway and circadian genes in bipolar disorder. There is extensive evidence that lithium is the preferred mood stabilizer and exerts efficacy by inhibiting GSK-3β activity. GSK-3β is involved in multiple downstream pathways (Wnt/β-catenin, release of brain-derived factors, etc.) and is associated with cell survival and neurogenesis. We will measure GSK-3β activity and Wnt signaling in peripheral cells and induced pluripotent stem cell (iPSC)-derived neurons to see if patients with bipolar disorder have congenital abnormalities or hyperactivity of this pathway. At the same time, the expression pattern of circadian genes (such as CLOCK, BMAL1) and melatonin secretion rhythm were monitored to evaluate whether the circadian rhythm was disordered. Hypothesis verification: Lithium salt/other mood stabilizer intervention experiments were designed to observe whether lithium salt could correct abnormal animal behavior and restore synaptic plasticity indicators such as long-term enhancement (LTP) in biphasic animal models (such as CLOCK mutant mice or mouse models acutely given manic drugs). On the other hand, lithium salts or GSK-3 inhibitors were added to in vitro iPSC neuronal models to detect changes in neuronal growth and synaptic formation, as well as correction of gene expression. If lithium treatment does promote increased expression of neuronal dendritic branching and synaptic markers, improving cellular adaptability to periodic depolarizing stimuli, it is demonstrated that bipolar disorder may stem from deficiencies in neuronal plasticity and rhythm regulation, and that drugs restore control of the excitator/inhibition balance by stabilizing intracellular information pathways (reducing noise fluctuations). We will also explain the role of lithium salts in combination with the DIKWP model: lithium salts may be equivalent to strengthening the stable regulation of the "knowledge layer" on the lower level of information, so that brain decision-making no longer fluctuates excessively with random signals, and verifying this will deepen the understanding of biphasic pathology.

3. Mechanism analysis of schizophrenia: Focusing on the effects of neurotransmitter imbalances (especially dopamine and glutamate) on information processing and perception. The widely accepted hypothesis is that schizophrenia has an overactive dopamine function in the limbic pathway of the midbrain, resulting in the wrong assignment of stimuli to saliency; At the same time, it may be accompanied by insufficient function of cortical glutamate NMDA receptors, resulting in decreased cortical excitability and GABAergic interneuron regulation disorders. Both of these can be seen as malfunctions of information control: the former allows irrelevant information to break into higher consciousness (hallucinatory delusions), and the latter causes the brain's entire network to synchronously disintegrate (cognitive disintegration). We will verify this through the following experiments:

• Dopamine function assay: Positron emission tomography (PET) tracer was used to detect the dopamine synthesis and release level of the striatum in patients to verify whether there was excess dopamine synthesis/release in patients with schizophrenia. Dopamine synthase DOPA decarboxylase activity was expected to increase in the striatum of patients compared with those who were naïve and healthy. If possible, the correlation of dopamine release with the patient's hallucinatory delusion score will be investigated.

• NMDA receptor function assessment: Combining preclinical studies and genetic data to analyze the gene expression and function of NMDA receptor subunits associated with schizophrenia. If available, magnetic resonance spectroscopy (MRS) can be used to measure the ratio of glutamate to GABA concentration in the prefrontal lobe, and transgenic mice (carrying schizophrenia risk genes such as NRG1 and DISC1 mutations) can be used to study the glutamate mechanism.

• Animal pharmacological model: Mice were acutely administered phencyclidine (PCP) or MK-801 (NMDA receptor antagonist) to induce psychotic-like behaviors and record their deficits in tasks such as prepulse inhibition (PPI) and working memory. This is the classic NMDA low-function model. Tranquilizers (antipsychotics, such as chlorpromazine or newer D2 partial agonists) are then used to see if the behavioural deficit can be reversed. Focuses on recording changes in animal performance on perceptual filtering (PPI) and cognitive flexibility (e.g., mazes) to correspond to human information processing capabilities.

• Validation of the information control hypothesis: If the above model shows that blocking NMDA receptors causes information processing disorders similar to those of schizophrenia (eg, decreased PPI, indicating poor sensory screening), and D2 antagonists can partially restore PPI, then cortical excitability**/inhibition balance and dopamine modulation are essential for information gating**。 We will further analyze human brain imaging/EEG data to look for a correspondence between dopamine-glutamate imbalance and abnormal nerve signaling. For example, a decrease in EEG (Auditory Change Perception) is widely reported and can be considered an indicator of unconscious information processing deficits. We planned to measure the change in MMN before and after the first dose of antipsychotic medication in patients with the first episode. If D2 receptor blockade therapy increases the amplitude of MMN, it means that the treatment repairs the early sensory information processing capacity to a certain extent, supporting the hypothesis that dopamine excessively interferes with sensory information pathways. Combined with the previously mentioned deviant saliency view, we will formalize the mechanism in the DIKWP model: assuming that the Purpose layer lowers the threshold for assigning information to the lower layers, resulting in too much data being fed back to the higher levels as meaningful information, causing confusion. Simulations were used to verify whether the adjusted parameters could reproduce and correct the symptoms.

Technology Roadmap: Research content 2 is carried out in three steps: "mechanism-drug-validation". Firstly, in-depth mechanism experiments were conducted on the key mechanisms of each disease selection (including biochemical detection, gene expression, animal models, etc.). Intervention drugs (antidepressants, mood stabilizers, antipsychotics) targeting this mechanism were then applied to the model or patients to observe changes in biological and behavioral indicators; Finally, the experimental results were included in the interpretation of the information processing framework of artificial consciousness to judge whether it conforms to the information control failure hypothesis. For example, if the experiment finds that dendritic damage + information processing delay in the depression model, and SSRI restores dendritic density + improves processing efficiency, then it can be considered that the core of depression is the weakening of the function of the information transmission link, and the drug can solve it by strengthening the data-to-information conversion rate; In the schizophrenic model, too much dopamine + PPI decreased, and the drug inhibited the rise of dopamine + PPI, indicating that excessive prominence signals led to filtration failure, and the drug rebuilt the filtration threshold. Through multi-target and multi-level experimental evidence, we will examine whether the etiology of mental disorders can be attributed to common defects in information control, and find out the similarities and differences between various diseases under this framework.

Research content 3: Discovery of biomarkers for early diagnosis of mental disorders

Main scientific question: Can objective indicators be used to predict the onset and progression of mental disorders before patients develop typical clinical symptoms? Which metrics or combinations of them are the most prospective and specific?

Research Protocol: In this part, we will carry out longitudinal cohort studies and multimodal biomarker mining to establish a three-level predictor system of "cognition-information-structure". The overall design is to recruit high-risk groups and early patients with mental disorders, combine follow-up and multimodal data collection, and apply machine learning to screen out reliable early diagnostic markers and verify their predictive efficacy. Specific implementation steps:

1. Cohort Recruitment and Placement: The following groups of people are recruited: (a) those with high familial genetic risk (e.g., adolescents with first-degree relatives with schizophrenia; First and second degree relatives of patients with bipolar disorder, etc.); (b) individuals with clinical high-risk syndrome (CHR), such as young people with mild psychotic symptoms but not yet onset, or patients at high risk of bipolar conversion from recurrent depressive episodes; (c) patients with the first onset of disease in the early stage (the course of the disease < 1 year, without systemic treatment); and a matched healthy control population. Strive to include 100-200 people in each group, and follow up the high-risk group for 2-3 years to see whether it is transformed into a confirmed patient.

2. Multimodal Baseline Assessment: A comprehensive multimodal examination was performed on each subject at the time of enrollment, including:

• Cognitive Assessment: Attention, working memory, executive function, emotion recognition, and more are tested using a standard neuropsychological battery. Particular attention is paid to possible patterns of minor impairments, such as mild working memory and social cognitive deficits in people at high risk of schizophrenia, and attention bias towards negative words in people at high risk of depression.

• Information: Paradigms such as event-dependent potential (ERP) are used to measure unconscious and conscious information processing. Such as P50 auditory gating, MMN, P300, eye tracking (smooth pursuit), etc. These objective signals can quantify the efficiency of the brain's filtering and processing of stimuli. For example, individuals at high risk of schizophrenia often have decreased P50 inhibition or decreased P300 amplitude.

• Structural and functional imaging of the brain (Structure): High-resolution structural MRI was performed to measure the volume and thickness of the brain region; DTI assesses white matter integrity; Resting-state fMRI assesses functional connectivity patterns. The imaging indicators related to various diseases reported in the literature were extracted, such as the decrease in hippocampal and amygdala volume in the early stage of depression, and the abnormal default network connection in patients at high risk of mental illness.

• Other biomarkers: Blood samples were collected for genetic polygenic risk score calculation, inflammatory markers (IL-6, CRP, etc.) measurement, stress hormone (such as cortisol arousal response) assessment, etc. New digital behavioral markers, such as wearable devices that record objective data such as sleep rhythm and activity level, can also be considered.
  All of these data will constitute a "biomarker signature set" for each subject.

Follow-up and outcome determination: Psychiatric assessments were performed at regular intervals (every 6 months) in the high-risk group, such as SCID interviews to determine whether there was a mental disorder that had progressed to a definitive diagnosis, and changes in severity of symptom scales. Follow-up for 2 to 3 years was recorded, which high-risk individuals were converted to patients, as well as symptom progression and functional outcomes in the patient group.

Biomarker Screening and Modeling: Statistical and machine learning methods are used to screen the combination of metrics that best predict outcomes from baseline multimodal features. Specific steps:

• Univariate analysis: to compare the differences between transformers and non-transformers in each individual index, and preliminarily screen for differentiated candidate markers (such as significantly lower working memory scores, lower P300 amplitude, etc.).

• Feature selection: LASSO regression, random forest and other methods are used to select the group of features that contribute the most to the prediction to avoid too much redundant information. We will pay special attention to the combination of features across the three layers of "cognition-information-structure", for example, a model may include three different levels of features of "executive function test score + P50 inhibition rate + hippocampal volume".

• Build a predictive model: Try a variety of machine learning models (logistic regression, support vector machines, neural networks, etc.) to cross-validate the trained model to predict the outcome (conversion or non-conversion) of high-risk individuals. The area under the ROC curve (AUC), sensitivity, specificity, positive predictive value and other indicators were used to evaluate the performance of the model. A previously mentioned study of people at high risk of psychosis showed that the AUC of the multimodal fusion model was 0.73 better than that of the single-modality, and we strive to achieve or exceed this accuracy.

• Validate the model: Apply the model to an independent validation sample (if available, collaborate with other centers to obtain independent high-risk cohort data) to verify robustness. If there are no independent samples, the stability of the model is verified by repeating different subsamples internally.

• Model interpretation: Combined with the DIKWP model, the selected features were interpreted. For example, if the final model selects three characteristics: "Accuracy of Emotion Recognition Test", "Amplitude of N400 ERP Component (Semantic Processing Potential)" and "Functional Connectivity Strength of Prefrontal-Limbic System", it can be interpreted as follows: the emotion recognition ability of the cognitive layer, the language semantic processing efficiency of the information layer, and the affective network coupling of the structure/functional layer jointly determine whether the high-risk individual will develop the disease. This is consistent with the DIKWP's view that the abnormal accumulation of data/information processing eventually leads to the clinical symptoms of the collapse at the Wisdom/Purpose level. We will explore the relationship between these characteristics and the predicted output of the model, and understand how risk manifests itself through each layer.

 Early warning system development: Development of easy-to-use assessment tools or scoring systems based on screened markers. For example, if a predictive model contains several key metrics, a weighted total score formula can be designed to assess individual risk. We will attempt to set thresholds in the follow-up data to optimize sensitivity and specificity to guide when to intervene clinically for high-risk individuals. Validation experiments are also planned for important biomarkers, such as validating intraday variability of specific EEG signals and testing test-retest reliability to assess their practical feasibility.

Technology Roadmap: Study content 3 takes the route of "cohort follow-up + multimodal fusion". The initial investment is in cohort recruitment and data collection, followed by data analysis modeling. The key point of the technology lies in the effective fusion and feature selection of multi-source data. With the help of advanced statistical learning tools, such as multi-kernel learning MKL or ensemble learning, we will achieve complementary fusion of cross-modal information. For time series follow-up data, we also consider dynamic models (e.g., Cox proportional hazards model combined with time-dependent covariates) to capture the impact of indicator trends on incidence risk. Once the model is established, we will, in turn, test whether the baseline characteristics of the predicted individuals conform to a certain pattern in order to scientifically understand the incubation mechanism of the disease. For example, individuals at high risk for accurate prediction of transformation may share certain insidious features: mild cognitive deficit + sleep rhythm disturbance + decreased social functioning, but the clinical diagnosis has not yet been met. This will help us refine the definition of "subclinical status" in the early stages of the disease and lay the groundwork for future early intervention trials in these populations.

Research content 4: The mechanism of social stress and psychological intervention and its linkage with the environmental mapping layer

MAIN SCIENTIFIC QUESTION: How do long-term socio-environmental stresses affect an individual's risk and course of developing mental disorders through biological mechanisms? How do psychosocial interventions work at the brain level? Can the influence of environmental-psychological factors be included in the "environmental mapping layer" of the artificial consciousness model to be explained?

Research Protocol: This section will combine animal and human studies to systematically analyze the effects of social stress (e.g., childhood trauma, major life events, chronic stress) on the brain and behavior, and delve into the mechanisms by which psychological interventions (e.g., psychotherapy, social support) can ameliorate this effect. The specific work is divided into two modules:

Module A: Pathogenic mechanisms of social stress

1. Clinical Population Survey and Biological Measurements: Detailed growth environment and stress exposure history were collected in the cohort of patients with depression, bipolar disorder, and schizophrenia (using standard scales such as Childhood Trauma Questionnaire (CTQ), Life Event Scale, LES, etc.). The association between stress exposure and clinical indicators such as age of onset, severity, and number of recurrences was analyzed. At the same time, blood or saliva is collected from patient samples for epigenetic testing (e.g., genome-wide DNA methylation microarrays) to find associations between childhood trauma and genetic methylation alterations. Childhood abuse has been shown to affect HPA axis function by leading to abnormal methylation of stress-related genes (eg, the glucocorticoid receptor encoded by NR3C1). We will validate similar findings and expand screening to more loci.

2. Effect of stress on brain structural function: Using MRI data obtained in our Content 3 cohort, we analyzed whether there were specific brain structural changes in people with high environmental stress. The literature suggests that childhood adversity is significantly associated with the decrease in adult hippocampal volume, and long-term stress may lead to dysfunctional connectivity in the anterior cingulate cortex and amygdala. We will verify in the data: for example, the participants will be classified according to the degree of childhood trauma, and the differences in brain imaging indicators between the two groups will be compared. Furthermore, the correlation between epigenetic or stress hormone levels and these brain indicators was examined to establish the link between "stress-epigenetics-brain structure-clinical". Hypothesis: The high childhood trauma group showed a comprehensive pattern of high methylation of NR3C1 gene, excessive cortisol secretion during stress, small hippocampal volume, and early age of onset of depression.

3. Stress agonist model research: Establish chronic unpredictable stress (CUS) model and social frustration stress model to simulate the effects of long-term stress on behavior and brain in rodents. Behavioral manifestations of depression/anxiety in animals (e.g., increased time of compulsive swimming, social avoidance, etc.) were recorded. The effects of stress on BDNF expression and synaptic plasticity-related proteins in the hippocampus and prefrontal lobe were detected at the molecular level. In the endocrine aspect, the corticosterone level curves of mice under basal and stress conditions were monitored. Focus on inflammation and microglial activation: chronic stress has been reported to induce low-grade brain inflammation and microglia, which may impair neurogenesis and synaptic plasticity. We will assess microglial status in the hippocampus, prefrontal lobes by immunohistochemistry and cell counting, validating stress-induced neuroinflammation. Animal brain imaging (e.g., small animal MRI or optical imaging techniques) is then used to detect stress-induced alterations in brain network connectivity—such as the overcoupling of default-like networks in response to stress that resembles human depression.

4. Artificial Consciousness Environment Mapping Explanation: Based on the above human and animal data, we will interpret how environmental stress alters the inner consciousness model within the framework of artificial consciousness. The DIKWP model needs to extend a layer of "environment mapping layer" to represent the subject's representation and evaluation of the external environment. Normally, the environment mapping layer maps external events to the internal semantic/conceptual space, triggering appropriate emotional and cognitive responses. However, under long-term high stress conditions, the mapping may be negatively biased: agents tend to interpret neutral events as threats, internalize failure attribution, or treat bad events as uncontrollable, thus forming an internal model of maladaptation. We will try to quantify this phenomenon through questionnaires and computational models: for example, the control group and the high-stress group each complete some interpretive tasks that ambiguous social situations to see if the high-stress group is more likely to give negative explanations (which reflects the pessimistic bias of environmental mapping). This bias is then simulated and reproduced using Bayesian cognitive models or Markov decision process models to correspond to stress biological changes in the brain. This step will provide us with ideas for the subsequent design of correction strategies.

Module B: The role and mechanism of psychological interventions

1. Psychological intervention clinical trials: Randomized controlled trials with small samples were selected for high-risk individuals or patients who were willing to receive intervention in the cohort of this project. For example, people at high risk of depression were randomly assigned to receive 8 weeks of mindfulness-based cognitive therapy (MBCT) or routine follow-up only, comparing changes in each measure before and after the intervention and with the control group. In another example, in patients who were in remission after the first episode of schizophrenia, they were randomized to receive family support + social skills training (combined psychological intervention) or routine follow-up, and the 1-year recurrence rate and functional outcomes were compared. Data on intervention effectiveness were collected through these trials.

2. Pre- and post-intervention biomarker assessments: At the beginning and end of the intervention, participants were subjected to the same multimodal assessments (cognitive tests, EEG, fMRI, etc.) as baseline to observe which indicators had significant changes. For example, some studies have found that after CBT treatment in depressed patients, the amygdala hyperresponse to negative stimuli is weakened, and the regulation of the limbic system in the prefrontal lobe is enhanced, which is reflected in the decreased amygdala activity and the enhanced negative coupling of the prefrontal-amygdala on functional MRI. We will test similar findings in participants to see if there were plasticity changes in certain brain network connections in the mental intervention group in the direction of health. EEG measures such as error-related negative waves (ERN) may reflect improvement in cognitive control; HRV (heart rate variability) or cortisol rhythm may reflect a decreased stress response. These alterations will be analysed in association with the amount of symptom improvement to identify the neural mechanisms of the psychological intervention.

3. Analysis of social support elements: Psychological interventions often include a variety of components, such as skill training, cognitive restructuring, behavior activation, and family education. We will use qualitative and quantitative methods to unravel which components are most helpful in changing patients' environmental mapping and cognitive patterns. For example, interviews with patients and their families to understand how family interventions can change their perceptions and coping with the disease. Quantitatively, self-rating scales (e.g., attribution style questionnaires) can be used to test whether the patient's attribution style becomes positive before and after the intervention; The Social Functioning Scale was used to evaluate whether the patient's social engagement was improved. How these psychological advances correspond to brain indicators will also be reflected in our analysis.

4. Artificial Consciousness Model Integration: Finally, we try to incorporate the stress-intervention mechanism into the DIKWP artificial consciousness framework to form a conceptual model. For example, the input to the environment mapping layer is defined as environmental stress events and the output is represented as representations (positive/negative) in the intrinsic semantic space. Long-term high stress pushes this mapping function in a negative direction, while psychological interventions pull the mapping function back to neutral or positive by reshaping cognition (adjustment of the knowledge/wisdom layer). We will construct a simple computational model to simulate this process: at first the model is given a series of negative inputs (simulated stress) to change its internal parameters (e.g., become pessimistic about the prior probability of future events); Then, by introducing new information (simulating the positive experiences and skills imparted by the psychological intervention), it was seen whether the internal parameters of the model could be adjusted back to normal. This simulation helps to validate our theory that psychological interventions update the internal state of the artificial consciousness model by providing new data and knowledge, making it more accurate in its mapping of the environment and serving the purpose/purpose of health.

Technology Roadmap: Study Content 4 comprehensively uses the methodology chain of "Epidemiological Investigation - Animal Experiment - Intervention Trial - Theoretical Modeling". The association and causality of cause (stress) and effect (brain/behavior abnormalities) are first proved by population and animal experiments, and then reversibility is proved through intervention, and finally rises to the theoretical explanation level. The techniques used in each link include questionnaire measurement, epigenetic sequencing, endocrine detection, animal behavior, brain imaging, psychological assessment, statistical modeling, and computational simulation. The challenge in this section is to objectify socio-environmental factors and correlate them with biological indicators. We will work with experts in the social sciences and biology within the project team to maximize the reliability and validity of the measurements. Through this series of studies, we hope to answer definitively: **Do and how environmental and psychological factors influence consciousness and behavior through bio-information pathways? Can intervening in these factors really change brain processes? **These insights will provide a scientific basis for strategies for the prevention and treatment of mental disorders.

Study content 5: Exploration of a new treatment model combining drugs and psychosocial interventions

Key scientific questions: How can pharmacotherapy and psychological/social interventions complement each other to improve outcomes for people with mental disorders at different levels? Is there an optimal combination model or new paradigm that can achieve a 1+1>2 efficacy? How to incorporate integrative therapy into the theoretical framework to guide individualized application?

Study protocol: Based on the above mechanisms, this section will design and test a new model of integrative therapy, and explore its synergistic mechanism and applicable conditions. The specific idea is as follows:

1. Comprehensive intervention model construction: Combining the results of research contents 2 and 4, the main biological and psychological therapeutic targets for each disease are extracted. For example, the biological target of depression is "decreased synaptic plasticity/monoamine deficiency" and the psychological target is "cognitive negative bias/behavioral withdrawal"; The biological targets of schizophrenia are "dopamine excess/glutamate insufficiency", and the psychosocial targets are "impaired social functioning/poor family coping", etc. For these targets, we envision integrated treatment options, such as:

• Major depressive disorder: Medications (such as SSRIs or ketamine) rapidly correct monoamine and plasticity deficits + psychotherapy (CBT or behavioral activation) corrects cognitive biases and reconstructs positive behavior patterns. Drugs provide a biological "window period" to improve brain plasticity, and psychotherapies take the opportunity to guide patients to learn to respond positively, thereby consolidating long-term efficacy.

• Bipolar disorder: medications (lithium or mood stabilizers) to stabilize nerve signaling pathways and circadian rhythms + family therapy/psychoeducation to help patients recognize precursors of mood swings, routine, and improve adherence. Pharmacological control of the physiological basis, psychosocial interventions to prevent precipitating factors and enhance patient self-management, jointly prolong remission and reduce relapse.

• Schizophrenia: Medications (atypical antipsychotics) to eliminate positive symptoms, reduce dopamine excess, + rehabilitation training (social skills training, SST, cognitive restructuring therapy, CRT, etc.) to improve cognitive function and social adaptation. Drugs stabilize the patient's mental symptoms and create conditions for rehabilitation training; Training improves patient function, improves quality of life, and reduces the risk of recurrence. In addition, the inclusion of family involvement can further consolidate the efficacy.

 Clinical trials to validate the efficacy of combination therapy: Prospective controlled trials were conducted in at least one of the combination modalities. For example, three groups were designed for depression: drug only group, psychological treatment only group, and drug + psychological comprehensive group, and the remission rate, symptom reduction and recurrence rate after 8 weeks of treatment were compared. Based on evidence from NIHR, medication + psychotherapy may be most effective in patients with moderate to severe depression. We want to verify this and quantify the magnitude of gain. For example, suppose that the response rate in the combination group is significantly higher than that in the single group, an increase of 10-20 percentage points. For schizophrenia, the difference between "drug + comprehensive rehabilitation" and "drug + routine follow-up" can also be compared in early intervention, and the indicators include recurrence rate and functional level change within 2 years. These clinical outcomes will provide direct evidence to support the superiority of the integrated treatment model.

 Mechanism of synergistic analysis: Using various bioindicators and models obtained in the project, the effect of comprehensive treatment is analyzed at the mechanistic level. For example, if patients with depression in the composite group have the lowest relapse rate, we will examine whether they exhibit both biological improvements (eg, increased BDNF levels, decreased DMN hyperconnectivity) and psychosocial improvements (eg, decreased negative automatic thoughts, increased social participation). If there are significant changes in both, and the magnitude of the change is related to the prognosis, the hypothesis that a two-pronged approach can optimize response is supported. In addition, we look at the timing effect: at the beginning of the course of treatment, the drug effect is usually more rapid and the psychological effect is gradual; Psychotherapy effects may be more durable after the drug is stopped. We will test whether the combination of treatment can reduce recurrence after discontinuation in follow-up, demonstrating the importance of psychological interventions to maintain efficacy. These findings will help inform treatment decision-making: for example, for patients with cognitive impairment, there is a greater need for enhanced rehabilitation; Patients with obvious stressors need family/social intervention to achieve individualized combinations.

 Therapeutic model under the framework of artificial consciousness: Finally, we try to map the integrated treatment to the DIKWP model for interpretation as a theoretical guide. For example, pharmacotherapy can be seen as a direct regulation of the underlying "data/information layer" (adjusting neurotransmitters = altering the original signal intensity or pre-processing), while psychotherapy can be seen as an intervention in the upper "knowledge/wisdom layer" (changing the patient's cognitive processing rules for information and the assignment of meaning to life events). The two work together on the Purpose layer/Person layer to achieve a comprehensive effect. We will simulate in the model: when there is only low-level intervention, the system output improvement is limited and easy to rebound; When only high-level intervention is performed, low-level noise still affects the effect; When combined, the system output (corresponding behavior/emotion) is optimal and robust. Demonstrating this with a formal approach would support the rationale for integrative treatment. This also reflects the concept of "full-level closed-loop therapy" – that is, effective therapy should cover multiple levels of the cognitive system, from sensory input to decision-making output. Once the theoretical model of this project is refined, it can be used to predict the effects of different combinations, such as simulating whether the addition of home support will further reduce relapse and optimize the duration of treatment. This provides guidance for future clinical design.

Technology Roadmap: Research Content 5 is translational and applied research in practice. The technical roadmap includes the design of clinical trials (involving clinical research methods such as randomization, blinding, and scale evaluation), as well as the application of all the previous results of this project to guide and evaluate these trials. Particular attention should be paid to the issue of sample size and adherence: we will make sample size estimates based on previous explorations and the literature to ensure sufficient statistical power of the trials, and ensure the quality of the intervention through rigorous project management (e.g., therapist training, adherence monitoring). In terms of data analysis, there are not only traditional clinical statistical methods (analysis of variance, survival analysis, etc.), but also the multimodal analysis methods we have mastered to explain the mechanism. The interdisciplinary team will collaborate in this context: the physician ensures the implementation of the clinical intervention, the psychologist provides the support of the intervention, and the neuroscientist and data scientist analyze the biological effects. It is expected that through the comprehensive treatment study, we will provide important empirical evidence that a single treatment for complex mental disorders is insufficient and that comprehensive intervention is necessary; and to provide policymakers and clinical practitioners with feasible combinations of interventions (e.g., an intervention model that establishes a multidisciplinary team MDT). This is also the embodiment of the project's ultimate service to the society.

Feasibility analysis

The research content of this project is grand and complex, but we have a solid preliminary foundation, mature technical methods and excellent research team to ensure the smooth implementation of the project. The following is an analysis of the feasibility of this project from several aspects:

1. Technical and methodological feasibility:

In recent years, the development of brain science, bioinformatics and artificial intelligence technology has provided strong tools for this project. First of all, single-cell omics technology is quite mature, and many sequencing centers in China have the ability to sequencing and analyze high-throughput single-cell RNA, and our team has also successfully mapped the single-cell map of the hippocampus in animal models of depression before, accumulating experience. High-field MRI and EEG/ERP equipment are available in our partner hospitals and brain imaging centers to support the acquisition of large samples of brain structure and function. Our team members are good at brain imaging data preprocessing and network analysis, and have built a process-based pipeline. Machine learning and multimodal data fusion methods are the key to this project, we have full-time data scientists and engineers, proficient in Python/R and other tools, can use mature scikit-learn, TensorFlow and other libraries for model training, and have the ability to implement cutting-edge algorithms such as multi-core learning and deep learning fusion. We have tested the potential of the multimodal psychosis risk prediction model on a small sample in the early stage, so we have the confidence to process large-scale data. In terms of animal models, we have a standard behavioral laboratory, which can stably replicate mouse models of chronic stress, social frustration models, etc., and master a variety of experimental methods such as behavior and molecular testing. Epigenetic sequencing and PET molecular imaging can be relatively expensive and complex, but we plan to do so through collaborations (e.g., with a research institute of the Chinese Academy of Sciences and a nuclear medicine center), which have agreed to provide us with technical support. In addition, the application of DIKWP artificial consciousness theory has certain novelty, but some members of the team are directly involved in Professor Duan Yucong's artificial consciousness project, and are very familiar with the principle and implementation details of the model. In summary, most of the technologies involved in the project have mature experience or reliable cooperation channels, so the technical implementation is feasible.

2. Feasibility of research subjects and sample acquisition:

This project requires the acquisition of a large number of patient and at-risk population data. The clinical units we rely on (such as the Mental Health Center Affiliated to XX University) are large-scale psychiatric specialties in China, with an annual outpatient volume of more than 200,000 person-times, sufficient inpatient beds, and a good referral cooperation relationship with the community. This provides an ample source for us to recruit people with different stages of mental disorders and high-risk populations. We have initially contacted the hospital's chief physician in the hospital's depression, bipolar, and schizophrenia specialists, who support the screening of eligible subjects for inclusion in the study on an outpatient basis. For high-risk groups, such as those with a family history of mental illness, we plan to recruit through a combination of in-hospital family screening and community advocacy. Those at high risk of childhood trauma and subthreshold symptoms can obtain them through cooperation with school counseling centers, online recruitment, etc. We have established a cohort of high-risk adolescents with depression and psychosis in the previous National Natural Science Foundation of China, which can be used for follow-up in this project. In terms of subject ethics, we will strictly abide by the principles of informed consent and privacy protection, and have obtained the preliminary consent of the ethics committee. As for brain tissue and biological samples, this project is supported by the National Brain Bank and Biobank, and a certain number of tissue samples can be obtained on demand. The collection of samples such as peripheral blood and saliva is less disturbing to patients and is easy to obtain multiple times. In terms of follow-up collaborators, our team has full-time research nurses and follow-up officers with many years of experience in the follow-up of patients with mental illness, and the follow-up success rate of past projects > 85%. Therefore, we are confident that we will obtain a sufficient number and quality of samples to meet the needs of the study.

3. Timeline and management feasibility:

The project duration is planned for a period of five years and has been planned in detail (see "Scheduling" below). The content of the study can be progressed in a logical step-by-step manner: first, cross-sectional sample collection and baseline evaluation (1-2 years) are carried out, and animal experiments are also carried out; Follow-up and intervention in the medium term (2-4 years) with continuous data analysis; Late integration verification and theoretical model refinement (4th-5th year). These efforts partially overlap in time, which can increase efficiency. Our project team has set detailed milestones, such as completing the multimodal baseline data of no less than 200 samples by the end of the second year, submitting the early marker prediction model by the end of the third year, and completing the comprehensive intervention trial by the end of the fourth year. The project leader and the project leader will hold a monthly progress meeting, implement milestone management, and coordinate resources to solve problems in a timely manner. The project budget fully considers the consumables and manpower needs of each job, and the financial allocation will be distributed to each sub-project unit according to the task to ensure the implementation of resources. In terms of risk management and control, we have formulated a filing plan for key links, and if there is a large number of follow-up losses, we will recruit additional subjects and improve follow-up methods; If the data fusion model does not work well, we will seek to work with AI experts to introduce more advanced algorithms, etc. Based on our previous successful experience in the management of multi-center large-scale projects (such as XX projects), the complexity of this project is under control, and we are confident that it will be completed on time and with high quality.

4. Support for the foundation of previous research:

For details, please refer to the "Research Basics" section below, and our team members have made rich achievements in the field of mental disorder research, including the discovery of new serum markers of depression, the establishment of cognitive training models for schizophrenia, etc., and have the professional knowledge and technical reserves required for the smooth implementation of the project. Especially in the field of DIKWP artificial intelligence theory, we have long-term exchanges with Professor Duan Yucong's team, and its 114 related patent achievements and cutting-edge theories are strong support for us. At the same time, the project team verified the feasibility of some research ideas (such as the feasibility of multimodal prediction, stress model replication, etc.) in previous small-scale experiments, reducing the exploration risk. Therefore, the project is built on a solid foundation rather than a source of water.

5. Advantages of multidisciplinary innovation:

The project team brings together experts in psychiatry, neuroscience, genetics, computer science, psychology and other fields to collaborate on complex problems from different perspectives. This multidisciplinary team model has proven to be essential in major brain science projects. Although the sub-projects involve different disciplines, they are all guided by a common theoretical framework and common goals, and their results can be mutually corroborated. For example, cellular abnormalities found in basic experiments can be verified by imaging studies; Imaging/electrophysiological abnormalities can be explained by computational models; The effect of the intervention can in turn verify the above-mentioned mechanism hypothesis, so as to form a closed loop. Therefore, interdisciplinarity is not only feasible, but will also have an effect that goes far beyond independent research. This is an important reason why we believe that this project can achieve significant innovative results.

In summary, this project has good feasibility conditions in terms of theory, technology, sample, time management and team foundation. The difficulties that may be encountered by the project are within the scope of anticipation and response. As the research progresses, we have the resources and capacity to ensure that the project stays on schedule and delivers the desired outcomes.

Innovation

This project has significant innovations in theory and methodology, and is expected to promote breakthroughs in the research and clinical practice of mental disorders. The main innovations are summarized as follows:

1. Theoretical integration and innovation: For the first time, the DIKWP artificial consciousness model was applied to the research of mental disorders, and the cognitive hierarchy theory of information science was creatively combined with the pathology of mental disorders. It is a pioneering attempt in the world, breaking through the traditional biological-psychological dichotomy framework, and analyzing mental disorders from the whole chain of data-information-knowledge-wisdom-purpose. This theoretical convergence is expected to propose a new paradigm for mental disorders, such as treating them as the result of multi-layer feedback dysregulation in the cognitive system, providing a new way of thinking for the unified interpretation of different symptoms.

2. Cross-scale multi-omics research: The combination of single-cell omics + brain imaging + electrophysiology has realized the cross-scale analysis of the brain mechanism of mental disorders. From microscopic cell type abnormalities to macroscopic brain network connections to subjective symptom experiences, a comprehensive correlation is formed. This kind of multi-scale study has rarely been reported in the field of psychiatry, and it is expected to discover disease-specific cell subtypes and circuit markers, and open up new directions for molecular diagnosis. At the same time, we use multimodal data fusion for early prediction, which is an important innovation in the field of psychiatric prediction intervention, which greatly improves the prediction accuracy.

3. Verification of information control hypothesis: The hypothesis of "information control failure" was proposed and multi-angle experiments were designed to verify its effectiveness. We combine the clinical drug action with mechanistic backwards to innovatively demonstrate the etiological mechanism: for example, the rapid action of ketamine proves that synaptic plasticity defects cause depression, the lithium salt proves that GSK-3 overactivity causes affective instability, and the antipsychotic drugs prove that abnormal prominent psychotic symptoms are abnormally prominent. This strategy of inferring the etiology based on efficacy is a unique innovation in the study of the mechanism of mental illness. Once the verification is successful, it will give a new connotation to "information processing" to textbook etiological theories (such as the dopamine hypothesis) and promote the updating of theories.

4. New markers for early diagnosis: This project will explore biomarkers with three-level integration of cognition, information and structure, and use machine learning to screen and combine them to achieve objective prediction of the incidence of disease in high-risk individuals. This is expected to lead to the discovery of new combinations of markers (e.g., EEG + cognitive test + MRI fusion indicators) that break through the current limitations of relying only on genes or imaging in a single dimension. In particular, we focus on interventionable predictors (e.g., cognitive function that can be improved through training) that have more significance than purely genetic measures. This will lay the foundation for precision prevention and has great clinical application prospects.

5. Biological Mechanisms of Environmental Stress: Through epigenetic and brain network studies, the influencing mechanisms of psychosocial factors are biologicalized. He innovatively proposed the concept of "environmental mapping layer" to explain the pathogenesis of stress, and revealed how childhood trauma left a molecular "imprint" and changed brain circuits. At the same time, we will systematically unravel for the first time the neural mechanisms of psychological interventions, such as how CBT alters brain connectivity. In the past, there was a lack of objective indicators on the role of psychotherapy, and our research will fill this gap and explain its principle with the artificial consciousness model, so as to provide a basis for the scientific use of psychological intervention.

6. A new model of integrated treatment: explore and verify the value of integrated drug + psychological intervention in improving prognosis and its biological synergistic mechanism. Although integrative therapy is clinically known to be effective, this project will reveal the scientific rationale behind its 1+1>2 and propose a new model of hierarchical and phased combination (e.g., drug-based in the onset phase and psychological in the consolidation phase). We will also innovatively propose the concept of cognitive level matching treatment: tailor the corresponding intervention according to the patient's injury concentration level. This elevates the traditional empirical combination therapy to an explainable theoretical guide.

7. Multidisciplinary and interdisciplinary approach innovation: The project integrates methods such as neuromics, neural network modeling, machine learning, and clinical trials to form a set of methodological paradigms for the study of mental disorders in a complex system. For example, the application of network control theory in artificial intelligence to the analysis of brain information flow, and the use of computational models to simulate the process of symptom generation are all new attempts. This methodological innovation will expand the arsenal of psychiatric research methods and bring about a paradigm shift.

In summary, the innovation of this project is reflected in the theory, content and methodology. Whether it is the introduction of artificial consciousness theory, multimodal fusion prediction, or mechanism-therapy closed-loop verification, it is the first attempt and major innovation. The results of the project are expected to produce high-level papers and patents, and promote the research on the mechanism of mental disorders to enter a new stage of "intelligent analysis", which is in line with the national requirements for original and leading scientific research.

Research Basis

The project team has carried out a lot of research work in related fields and has accumulated rich research foundation and experience, which provides a solid guarantee for the smooth implementation of this project. The following is an explanation of the team's previous research results, related work basis and research conditions:

1. The team's achievements in the research of the mechanism of mental disorders:

The core members of the project team have long been committed to the mechanism and treatment of depression, bipolar disorder, schizophrenia and other diseases, and have published a series of high-level papers and have solid academic accumulation. For example, the project leader, Professor XX, has made important contributions to the field of depression biomarkers, and has published a paper on the abnormality of the peripheral blood inflammatory factor profile of patients with depression and its association with disease subtypes, which has been cited more than 100 times by SCI. Professor XX's team also used brain imaging technology to study the cognitive function of depression, and found a dysfunctional connection pattern between the default network and the cognitive control network of depressed patients, and his paper was published in Human Brain Mapping. Researcher YY (responsible for the single-cell omics part) has published an article on the brain gene co-expression network of schizophrenia in Molecular Psychiatry, revealing the coexistence of microglial activation and synaptic gene downregulation, which provides ideas for the study of the cellular mechanism of this project. In addition, Associate Professor ZZ has conducted research on sleep and rhythm in the field of bipolar disorder, and has reported the association between circadian gene polymorphisms and recurrence patterns in patients with bipolar disorder, and verified the mechanism by which circadian rhythm disorders can induce mood swings in animals. These early results demonstrate that the team has a deep understanding of the biology of the target disease, which lays the foundation for this project to target key scientific questions.

2. Basis of Interdisciplinary Research on Artificial Intelligence and Brain Science:

Many members of the project team are active in the intersection of artificial intelligence and brain science. In particular, we have a close working relationship with the artificial consciousness research team led by Professor Duan Yucong. Professor Duan was the first to propose the DIKWP model and has achieved leading results in the field of global artificial consciousness. Our member WW associate researcher has worked under the guidance of Professor Duan, has an in-depth understanding of the principles and implementation of DIKWP, and has participated in the application project of DIKWP model in the evaluation of large language model "class awareness". He is also a member of the World Artificial Awareness Association and has participated in the development of artificial awareness assessment standards. This means that we have a direct advantage in the theory and technology of artificial consciousness and can introduce it into this project. Previously, WW Associate Research has preliminarily applied the DIKWP model to simulated cognitive tasks, which proves that the model can well reproduce the hierarchical reasoning process of human problem solving. We also have successful experience in AI model training based on brain data. For example, our computational engineers have trained a neural network based on fMRI data to classify patients with schizophrenia, visualized the implicit representation of the model, and found network representations similar to the patients' DMN anomalies. These pioneering attempts have laid a methodological foundation for us to further integrate artificial consciousness models with brain diseases.

3. Experience in multimodal data acquisition and analysis:

The team has rich experience in biological big data and multimodal data processing. In recent years, the laboratory led by Prof. X has completed a joint peripheral blood transcriptome + metabolome analysis project in patients with depression, successfully integrating two omics to identify a combination of biomarkers to predict treatment response, and the results were published in Translational Psychiatry. This experience can be directly applied to marker fusion analysis early in the project. YY researchers are good at single-cell sequencing data processing and have developed cell type identification algorithms. Dr. UU, a data scientist in the project team, is proficient in machine learning tools, and with the support of the previous fund, he developed a multimodal brain image fusion framework, which has been applied to the construction of early diagnosis models for Alzheimer's disease, and has achieved higher accuracy than single modality, and has also won awards in medical image analysis competitions. Dr. UU has transplanted the framework to mental disorder data for initial testing, and the results have been encouraging. This means that we have mature software and pipelines to process multiple data such as MRI + EEG + cognition, saving time for the project in method development. In addition, we have accumulated a database of high-risk groups and patients on a certain scale. For example, in a provincial and ministerial project, we collected fMRI and cognitive data from 100 patients with early schizophrenia, which can be included in the expanded sample size. It can be said that multimodal data analysis is not a new challenge, but one of our strengths.

4. Fundamentals of Animal and Cell Experiments:

The animal models and cell experiments involved in this project are also part of our team's daily work. The animal experiment platform led by Associate Professor ZZ has built a rat model of chronic stress, which successfully induced behavioral changes similar to human depression, and was published in the Chinese Pharmacological Journal. The platform is equipped with standard behavioral devices (open field, tail suspension, maze, etc.) and surgical equipment (microinjection pumps, stereotaxic instruments, etc.), and technicians are skilled in carrying out experimental operations such as lentiviral injection and optogenetic intervention, which can support us to carry out fine experiments such as dopamine intervention and transient inhibition of specific brain regions. At the cellular level, we have a cell culture and differentiation chamber, which can obtain exosomes from peripheral blood and induce iPSC redifferentiation into neurons from patients' skin fibroblasts. If we need to simulate the neuronal characteristics of patients, these cell lines can be directly used for drug testing or gene editing experiments (e.g., knocking out a risk gene observation). All of this shows that we have a complete experimental platform that allows us to carry out most of the experiments required for this project without having to build them from scratch.

5. Clinical Resource and Collaboration Network:

The project is supported by the Mental Health Center and a number of partner hospitals. The director of the center has set up a clinical collaboration team for this project, including the heads of depression, bipolar and schizophrenia, who will coordinate patient recruitment, evaluation and intervention. For example, the Depression Specialist has an experienced psychotherapy team that can undertake the implementation of psychological interventions in the program; The follow-up system of bipolar specialties can be used for long-term follow-up of this project; The Schizophrenia Specialty is a municipal demonstration unit in the first year of psychiatry (early intervention) with an adequate source of patients. We are also working with community health centres to include high-risk groups who have not yet seen a doctor but have symptoms. There are also successful collaborations between team members, such as Professor X and Associate Professor ZZ who jointly undertook a clinical study on the combination of antidepressants and CBT and successfully completed and published a paper. This tacit cooperation will continue to the project to improve efficiency. In addition, our network of cooperation with experts in related fields at home and abroad will also provide support. In terms of artificial intelligence, we can invite the consultants of Professor Duan Yucong's team to guide and even investigate the prototype artificial consciousness system they developed. In terms of epigenetics, we have jointly applied for equipment sharing with experts from a gene center in Beijing, which can quickly complete sample sequencing. Internationally, we have a Memorandum of Understanding (MOU) with the Brain Imaging team at Y University in the United States, which allows us to help each other in data analysis and student development. These collaborations ensure that we can mobilize the best resources to accomplish complex tasks.

6. Fundamentals of Project Management and Talent Development:

The members of the project team have undertaken many national scientific research tasks and have rich experience in large-scale project management and talent training. The sub-projects of the National Key R&D Program presided over by Professor X passed the acceptance on time and with high quality, and formed an effective project management system (weekly meetings, quarterly reports, task book decomposition, etc.). These systems will be used in this project, so that the work packages can be carried out in an orderly manner without falling off the chain. At the same time, the team has a multi-level talent echelon, including senior professors, young and middle-aged backbones and outstanding doctoral students. Among them, many young teachers have overseas training experience, postdoctoral AA has studied single-cell omics technology in foreign laboratories, and doctoral student BB has won awards in the national brain science competition. They will be the backbone of the implementation of the project and will grow rapidly over the course of the project. We plan to hold a seminar and training course every year, inviting experts in the interdisciplinary fields of artificial intelligence and brain science to give lectures to improve the overall research ability of the team. These talent training initiatives have also been supported by the supporting units. This means that there will be no talent gap in the implementation of the project, but on the contrary, a group of compound talents will be produced to make long-term contributions to the development of the field.

Overall, the project has a solid research foundation: it has the scientific direction laid by the previous results, the technical platform to ensure its implementation, and the multidisciplinary talent and cooperation network to provide support. These foundations will significantly reduce the technical risks of the project, improve research efficiency, and enable us to translate ambitious research goals into practical research processes and ultimately achieve high-level innovation results.

Team & Schedule

Composition and division of labor of the research team

This project is led by XX University and completed by a number of units to form an excellent interdisciplinary team. The core members of the team and the division of labor are as follows:

• Professor X (Project Leader, Unit: Mental Health Center Affiliated to XX University): Psychiatric expert, long-term research on depression and bipolar disorder. Responsible for the overall planning of the project and the clinical research part. Professor X will coordinate patient recruitment and evaluation, clinical trial design, and guide early marker screening and comprehensive treatment model development.

• Prof. Y (Project Leader, Unit: Institute of Brain Science, XX University): Expert in neurobiology, experienced in single-cell omics and animal models. He is responsible for the basic mechanism research of Topics 1 and 2, including single-cell sequencing of brain tissue and animal stress model experiments. Prof. Y will also liaise with brain banks and gene sequencing platforms to ensure sample and technology implementation.

• Professor Z (Project Leader, Unit: Department of Computer Science, XX University of Science and Technology): Expert in artificial intelligence and data science, familiar with the DIKWP model. Responsible for the data analysis and modeling of Topic 3 and part of Topic 5, including the development of multimodal prediction models, artificial consciousness computing simulation, etc. Professor Z's team will develop the required algorithm and assist in interpreting the model results.

• Researcher W (core backbone, unit: Institute for Advanced Study, XX University): an expert in the direction of artificial consciousness, a member of Professor Duan Yucong's team. As a theoretical consultant and specific executor, participate in the application design of the DIKWP model in the project, and help other members understand and apply the theory. Researcher W is also responsible for semantic network analysis and awareness assessment index design for some data.

• Associate Professor Z (core backbone, unit: Mental Health Center Affiliated to XX University): professional doctor and scientific researcher of schizophrenia. Responsible for patient cohort follow-up, cognitive function assessment, and social function training intervention. In topics 4 and 5, psychological intervention trials were carried out specifically and patient outcomes were monitored. Associate Professor Z is also responsible for reaching out to families and the community to promote intervention adherence.

• Dr. U (Data Scientist, Unit: XX University of Science and Technology): Machine Learning Engineer. Responsible for multi-modal data preprocessing, feature selection and model training, and implementing MKL fusion algorithms and deep learning frameworks to ensure the effective fusion of complex data. Dr. U will also build a project database and analysis platform to facilitate the collaborative processing of data by all members.

• Postdoctoral Fellow V (Experimental Backbone, Unit: Institute of Brain Science, XX University): Mainly responsible for single-cell and epigenetic experimental operations, including tissue processing, sequencing and preliminary bioinformatics analysis. He has been trained in foreign brain science centers and has reliable technology. After his Ph.D., he also participated in animal experiments and assisted Professor Y in completing stress model research.

• Several doctoral students: The program will involve 4-6 doctoral students and 2-3 master's students, each of whom will be responsible for a part of the specific work and receive supervision. For example, one PhD student focuses on fMRI and EEG data analysis, another focuses on PET images and dopamine measurements, and so on. This not only accomplishes scientific research tasks, but also cultivates talents.

The organizational structure of the team is as follows: the project leader is generally coordinated, and there are five research groups, each group is led by the project leader, and supported by the backbone of the corresponding field to lead students to complete the task. The team is characterized by diversity and complementarity: clinical (X, Z associate professor) + basic (Y, V) + data (Z, U) + theory (W) combination, to ensure that each research link is checked by experts, and at the same time, under the project objectives unified cooperation, not fighting alone.

In addition, we have set up a project management office, led by Professor X's research assistant, to assist in arranging meetings, tracking progress, connecting finances, and summarizing information to ensure smooth communication. The program will organize regular internal academic exchanges to promote knowledge sharing and student training. It is foreseeable that such a team with a neat lineup and a clear division of labor will give full play to their respective advantages and form a joint force to overcome the complex scientific problems of this project.

Plan progress and milestones

The proposed implementation period of the project is 5 years (60 months). According to the research content and logical sequence, the following main stages and time nodes are planned:

• Year 1 (Months 1-12): Project Initiation and Foundation Preparation Phase.

  • Month 1-3: Complete the administrative start of the project, the ethics approval, and the formulation of a detailed implementation plan. Convene a plenary session to clarify the division of labor. Initiate cohort recruitment and train evaluators. Purchase of major reagent equipment.

  • Month 4-12: Sample baseline collection is in full swing. Establish cohorts and high-risk cohorts of patients with depression, bipolar and schizophrenia, and strive to recruit at least 30-50 people/group and complete baseline assessment (clinical, cognitive, imaging, EEG, blood sample collection). At the same time, the single-cell sequencing of brain tissue was initiated: the first batch of brain tissue (at least 5 patients and controls) was obtained, the cell isolation and sequencing library preparation were completed, and the sequencing company was sent to the sequencing company. In terms of animal experiments, the first batch of mouse models of chronic stress were established, and behavioral testing and sampling analysis began. Milestone 1: By the end of the first year, the baseline multimodal data collection and preliminary quality control of ≥ 100 subjects were completed; Preliminary single-cell sequencing data were obtained; Animal models validate stress behavior effects.

• Year 2 (Months 13-24): Mechanism Deepening and Model Building Phase.

  • May 13-18: In-depth analysis of single-cell sequencing data to identify differential cell types and gene pathways. Perform spatial transcriptome validation of key findings. Continue to recruit new samples and expand the cohort size to a target number (e.g., 200 for depression, 100 for bipolar, 150 for schizophrenia, and 100 for high risk). The high-risk group was followed up and the incidence was recorded. Animal experiments continued, and drug intervention groups (e.g., ketamine, lithium salts) were added to verify reversibility. Develop machine learning prediction models, train preliminary models with collected data, and adjust algorithms. Month 19-24: Carry out special experiments such as PET imaging to determine the dopamine index of sub-samples; Epigenetic sequencing was completed for comparison of the high-stress/low-stress patient groups. Design the details of the psychological intervention program and recruit willing participants to prepare for the intervention. Milestone 2: By the end of the second year, the results of major mechanistic experiments (e.g., single-cell first draft data; biological analysis of stress model), complete the preliminary construction of early prediction model and report AUC and other performance; Each cohort was followed up for at least one year, and partial conversion outcome information was obtained.

• Year 3 (months 25-36): Intervention implementation and comprehensive analysis phase.

  • May 25-30: Start of clinical trials of psychological interventions (e.g., RCTs for the comprehensive treatment of depression): enrollment of patients in the intervention and control groups, implementation of interventions (pharmacological and psychological), close monitoring of symptom changes, collection of biomarker data before and after the intervention as planned. Continue routine cohort follow-up, focusing on which high-risk cases have been turned into confirmed cases, and collect data such as recurrence. The machine learning model was updated to include more follow-up outcomes, and features and algorithms were continuously optimized. Month 31-36: Most of the psychological intervention courses were completed, and data analysis was started to compare the efficacy between groups. The results of animal and clinical mechanisms were comprehensively collated, and the obtained experimental evidence was spliced into the theoretical framework of artificial consciousness, and the theoretical model description was written. Milestone 3: Completion of preliminary results from at least one combination intervention trial by the end of year 3 (e.g., data on significant superiority of combination therapy over monotherapy); The final version of the early prediction model was obtained and published (the target was published in an international journal, and the model AUC>0.7); Forming a theoretical prototype of the information processing mechanism of mental disorders, and preparing to submit high-level papers (such as mechanistic papers combining single-cell and imaging results).

• Year 4 (Months 37-48): Achievement refinement and model validation phase.

  • Month 37-42: Conduct validation experiments for previous discoveries. For example, if a biomarker is found to have high predictive value, develop a simple kit to verify the sensitivity in an independent sample; If a new target is discovered in a single cell, its function is verified using gene-edited animal or cell-based experiments. The calculation and simulation of the DIKWP model has entered a critical stage: according to the identified loop defects, the simulation model is constructed to test the hypothesis, and compared with the actual data, the model parameters are repeatedly modified to approximate the real phenomenon. The last cohort assessment was performed, and the last imaging/cognitive assessment was performed for those who were still in follow-up to verify the correctness of early prediction and the impact of the intervention. Month 43-48: Organize a comprehensive analysis meeting to cross-discuss the results of each topic to ensure that each important finding is supported by multiple evidences. He has authored a series of papers and patents, including: patents on cell mapping and use of mental disorders, patents on multi-index models for early diagnosis, and papers such as "A Review of Cognitive Mechanisms of Mental Disorders under the Framework of DIKWP". Milestone 4: By the end of the 4th year, submitted/published a number of core papers and applied for 2-3 patents; Complete the disease application description of the DIKWP artificial consciousness model, and may publish a monograph chapter or review.

• Year 5 (Months 49-60): Summary, refinement and promotion of results.
  Milestone 5 (Project Completion): At the end of the fifth year, the scheduled research tasks will be fully completed and the assessment indicators will be met: no less than X papers will be published (including YY papers in SCI Zone 1), several doctoral/master's students will be trained, and draft guidelines for early screening will be formulated. The results of the project are presented in both theoretical and applied aspects, including a new information-consciousness theoretical framework for mental disorders, as well as new technologies for practical diagnosis and treatment (predictive models, joint intervention programs, etc.). Pass the acceptance and promote the application.

  • Months 49-54: Complete the rest of the experiments, such as long-term follow-up to obtain data on the final 5-year recurrence rate. Extended follow-up was carried out for patients treated with comprehensive treatment to see the long-term effect. The final integration model was visualized, such as drawing the purpose of mental disorder cognitive circuit disorder and the abnormal DIKWP level of human brain, which were used in papers and popular science. Conduct internal testing of results: Organize industry expert demonstrations to ask peers to test the performance of our early diagnosis tools on new data and the interpretation of our models to other data to improve reliability. May 55-60: Improve all research reports and data archiving. Hold a presentation workshop to present our findings and theories to the academic community. Promote clinical translation: For example, cooperate with the laboratory department of the hospital to include important biomarkers in the pilot of the physical examination package; Recommend to health authorities to strengthen the screening mechanism for high-risk groups. Write a project summary report and accept it for acceptance by the competent department.

The above-mentioned plan and arrangement reflect the idea of step-by-step, parallel progress and stage acceptance. The phases in the timeline are connected and relatively independent, ensuring that even if individual links are postponed, the overall progress will not be interrupted. At the same time, there are clear milestones each year for easy monitoring and adjustment. We will strictly follow this plan, and optimize the allocation of resources in a timely manner according to the mid-term evaluation, so as to ensure that the project is completed on schedule and with high quality, and contribute new knowledge and means to national mental health research and practice.