summary

Neurodegenerative diseases associated with movement disorders (such as Parkinson's disease, amyotrophic lateral sclerosis, ALS, Huntington's disease, hereditary cerebellar ataxia, etc.) pose a serious threat to human health. These disorders are usually caused by abnormal aggregation of specific pathological proteins and neurologic degeneration, and patients present with progressive motor dysfunction, often accompanied by cognitive and behavioral abnormalities. At present, there is no cure for this disease. Based on the "Data-Information-Knowledge-Wisdom-Purpose" artificial consciousness theoretical model proposed by Professor Duan Yucong, this project aims to reveal the pathogenic mechanism of neurodegenerative diseases such as movement disorders from multiple levels and explore innovative treatment strategies.

In this study, we will comprehensively use molecular biology, cell and organoid models, animal models, and human genetic data to focus on three aspects: (1) analyze the abnormal folding aggregation and cross-cell transmission mechanism of core pathological proteins in Parkinson's disease and other diseases (such as α-synuclein, TDP-43, HTT, etc.), and use the DIKWP model to analyze the association between data perception, information processing and uncontrolled purpose in this process; (2) Based on the theory of artificial consciousness and the hierarchical control framework of DIKWP, brain-like organoids and novel mammalian disease models were established to simulate the process of disease progression, focusing on the interaction between "brain cognitive control layer disintegration" and neural network regression. (3) Combining multi-omics and network pharmacology methods, we can find potential new targets for clinical intervention and design intervention strategies, and construct a DIKWP-drug regulatory pathway mapping model to simulate the transmission chain of pharmacodynamic effects, so as to guide the development of new treatment regimens.

The project is characterized by introducing the cutting-edge DIKWP artificial consciousness model in the field of artificial intelligence into the research of neurodegenerative diseases, understanding the mechanism of disease occurrence and development through the five-level integrated perspective of "data-information-knowledge-wisdom-purpose", and bridging the gap between molecular pathological changes and cognitive dysfunction. With the support of comprehensive and in-depth basic research, this project is expected to elucidate the key pathogenic links of neurodegenerative diseases such as movement disorders, propose new intervention targets and strategies, and establish theoretical models that can be used for drug screening and efficacy prediction, which has important scientific significance and application value for overcoming the clinical problems of such diseases.

Basis for the project

1. Background and significance

Neurodegenerative diseases associated with movement disorders include Parkinson's disease (PD), amyotrophic lateral sclerosis (ALS), Huntington's disease (HD), hereditary cerebellar ataxia (SCA), etc. These disorders are characterized by progressive deterioration of motor function, often accompanied by abnormalities in cognitive function and behavior, and place a heavy burden on the patient's family and society. Epidemiological data show that as the population ages, the number of people with diseases such as PD will rise significantly. For example, the 2025 British Medical Journal (BMJ) research model predicts that there will be more than 25.2 million Parkinson's disease patients worldwide by 2050 people, an increase of about 112% compared to 2021. Although ALS is a rare disease, the number of ALS patients worldwide is expected to increase from about 223,000 in 2015 to about 377,000 in 2040 (an increase of 69%) due to its high mortality and significant societal impact. The median survival of patients with ALS was only 24 years; Hereditary disorders such as HD and SCA tend to develop in middle age, with severe movement impairment and cognitive deterioration occurring within 1020 years. At present, there is no fundamental treatment for these diseases: Parkinson's disease can only improve symptoms through measures such as dopamine replacement, but cannot stop the process of neurodegeneration; In ALS, only drugs such as riluzole and edaravone can slightly delay the course of the disease, but the disease is irreversible; There is no treatment for Huntington's disease and most inherited ataxia that can alter the course of the disease. Therefore, elucidating the pathogenesis of these diseases and exploring novel intervention strategies is an urgent challenge for neuroscience and clinical medicine.

2. Existing research foundation and shortcomings

**Common pathological features: aberrant protein aggregation and neural network degeneration. **The clinical manifestations of different degenerative diseases vary but there are commonalities at the pathological level: abnormal aggregation and deposition of specific proteins in the brain or spinal cord, and consequent neuronal loss and network dysfunction. Lewy bodies accumulating α-synucle α in in the substantia nigra dense neurons in patients with Parkinson's disease; TDP-43 protein deposition is common in motor neurons in ALS patients; Striatum neurons in HD patients are filled with polyglutamine-expanding huntingtin (HTT) aggregates. Hereditary cerebellar ataxia (e.g., SCA1/2/3, etc.) is also caused by mutations and accumulation of specific proteins (e.g., ataxin-1, ataxin-3, etc.). These misfolded proteins not only play the role of "core pathological proteins" in their respective diseases, but their aggregation behavior also presents ** "seed transmission" or "prion transmission".Properties: that is, the initially formed abnormal conformational protein can be used as a template to induce a conformational transition of normal proteins, thereby propagating pathology along neural circuits. For example, the α-synuclein pathology of Parkinson's disease has been found to spread progressively along the vagus and autonomic nerves to the midbrain and cerebral cortex, supporting its "prion-like" spread in the brain. Similarly, TDP-43 pathology in ALS has been shown to spread through neuroanatomical junctions, and TDP-43 has been observed to spread along the corticospinal tract in models, causing ALS-like symptoms. Huntingtin aggregates may also metastasize across neurons via synaptic vesicles, contributing to disease progression in the brain. These findings suggest aberrant dynamics of protein misfolding and aggregation, as well as cascading propagation in neural networks, as well as a common pathogenic mechanism for dyskinesia degenerative diseases. Traditional studies have characterized protein aggregation and toxicity at the molecular and cellular levels, but there is still a lack of understanding of how multi-level information processing dysregulation leads to global neurological collapse**. For example, with the same protein aggregation, why does the pattern of symptom progression and neural circuit involvement differ in different patients? What are the internal "control mechanisms" of the nervous system that try to cope with the abnormal protein load, and how can their failure lead to dysfunction? These issues have not yet been fully elucidated.

The need for multi-level research and integration: The research on such diseases is moving towards "cross-scale and multi-dimensional". On the one hand, with the development of high-throughput omics and imaging, a large number of data at different levels (genes-proteins-cells-circuits-behaviors) have emerged. On the other hand, novel models such as stem cell-derived brain organoids and more sophisticated animal models have made it possible to observe disease progression in vivo. However, it remains difficult to integrate multi-layered information into a holistic understanding of the disease. At present, most studies and models can only cover part of the scale: in vitro cell and organoid models are able to reproduce molecular and cellular lesions, but lack the in vivo environment and advanced functions; Animal models can demonstrate system-level symptoms, but often species differences limit the extrapolation of results. In addition, the lack of simulation of cognitive and network functions is also a limitation of the existing models. For example, traditional brain organoids are unable to form higher-order functional networks and cognitive outputs due to the lack of sensory input and environmental interaction. This makes it difficult to rely solely on organoids to study the phenomenon of "cognitive control collapse"** in patients (e.g., cognitive executive dysfunction evident in late Parkinson's and Huntington's disease). Similarly, animal models have difficulty assessing interactions between higher brain functions and neurodegeneration due to limited cognitive abilities. In summary, there is an urgent need for a theoretical framework that can penetrate multiple levels from molecular to cognitive, integrate multidimensional data to explain disease mechanisms, and guide the design of new models and therapies.

3. Introduction of the DIKWP model and artificial consciousness theory

In order to solve the above problems, we propose to introduce the DIKWP model and artificial consciousness (AC) theory in the field of artificial intelligence into the study of the mechanism of neurodegenerative diseases. The DIKWP model is a new type of cognitive model developed by Professor Duan Yucong's team, which introduces a higher level of "Purpose/Intention" on the basis of the classic data-information-knowledge-wisdom (DIKW) pyramid to form a five-level cognitive system. The model adopts a network interaction structure, so that the semantic information at all levels can be fed back and updated in both directions. In short, DIKWP consists of five levels of bottom-up data, information, knowledge, Wisdom and the highest Purpose/Purpose, from the bottom level of the perception and processing of the original data, gradually extracting information and knowledge, rising to the decision-making of the Wisdom layer, and finally guided and constrained by the topmost Purpose layer. This expansion emphasizes the driving role of decision-making purpose and subjective purpose in the cognitive process, and combines subjective purpose with objective information processing, which is regarded as a major innovation in the field of artificial intelligence cognitive computing. Under this model, every step of the AI system's decisions can be explained retrospectively, because there is a clear purpose to guide it.

Complementing the DIKWP model, the artificial consciousness theory aims to equip AI with human-like consciousness characteristics, including advanced functions such as self-monitoring and autonomous purpose. The DIKWP artificial consciousness model proposed by Professor Duan Yucong's team is based on information processing and purpose-driven mechanism, which realizes the simulation of consciousness generation and function through the transformation and interaction of data, information, knowledge, wisdom and purpose. Specifically, the team has tried to combine the DIKWP model with neuroscience concepts to establish the "DIKWP theory of brain region mapping", which simulates the generation and flow process of data-information-knowledge-wisdom-purpose in the brain, and builds a deep cognitive model. The researchers mapped the objective external performance and subjective internal cognition into the data interaction and reasoning process within the model, and built a prototype system of DIKWP physiological and artificial consciousness to verify the feasibility of the framework. In short, the DIKWP Artificial Consciousness Framework provides a multi-layered unified cognitive control model that can be used to describe how the brain moves from low-level signal processing to high-level purpose-directed behavior.

We believe that the application of the DIKWP model to the study of neurodegenerative diseases has the following potential advantages:

• Multi-scale integration capability: The DIKWP framework naturally contains a hierarchical mapping from "data" to "wisdom / purpose", which is similar to the macro function of associating objective data at the molecular and cellular levels all the way to the cognitive and behavioral levels. This is highly consistent with the requirements of modern neuroscience for cross-scale mechanisms. This model is expected to serve as a "bridge" to understand the experimental results at different levels under the same semantic system.

• Emphasis on purpose-driven and control dysregulation: Many neurodegenerative symptoms (e.g., Parkinson's disease, movement disorders, cognitive executive dysfunction, etc.) can be seen as dysregulation of higher control functions in the brain. The "Purpose" layer, which is particularly emphasized in the DIKWP model, represents the supreme authority of decision-making and control. This framework allows us to characterize the impact of the disintegration of the purpose layer on the lower level in the disease state, and to reinterpret the causes of movement disorders from the perspective of information processing. For example, in patients with Parkinson's disease, the regulation of motor circuits at the purpose layer is weakened due to the lack of dopamine in the brain, and abnormal discharge (tremor, etc.) occurs in the low-level motor information processing, which can be explained in the DIKWP model as a situation in which the loss of the top-level control signal leads to the increase of the lower-level information noise.

• Artificial Consciousness Simulation Helps New Models: Using the artificial consciousness theory, traditional biological models can be given certain "cognitive" elements and expand their experimental capabilities. For example, brain organoids can be subjected to controls or inputs from artificial intelligence that allow them to interact with their environment, thereby partially simulating sensory-motor circuits and learning processes. Recent studies have demonstrated that cultured human brain organoid/neuronal networks exhibit goal-directed learning when given environmental feedback (e.g., human brain cells in a petri dish learn to play the game of ping-pong under electrical stimulation-feedback conditions), showing the potential to generate cognitive function in an in vitro system. This provides important implications for us to improve the disease model based on AC theory.

In conclusion, the DIKWP model and artificial consciousness theory provide a new perspective for understanding and simulating neurodegenerative diseases. It is expected to correlate molecular pathology with systemic cognitive function, help explain complex pathological transmission and dysfunctional mechanisms, and guide us in designing novel disease models and intervention strategies that include a "cognitive control layer". Based on this innovative idea, this project intends to use the DIKWP artificial consciousness model as a unified theoretical framework, integrate experimental and computational methods, systematically study the pathogenic mechanism of neurodegenerative diseases such as movement disorders, fill the lack of multi-level integration of existing research, and help find a breakthrough in treatment.

Research Objectives:

The overall goal of this project is to use the DIKWP model and artificial consciousness theory to comprehensively analyze the pathogenic mechanism of neurodegenerative diseases such as movement disorders from the molecular to the systemic level, and then propose a verifiable new intervention strategy. Specific objectives include:

1. Elucidate the pathogenic mechanism of aggregation and propagation of key pathological proteins: Using cells, brain-like organoids and animal models, combined with patient genetic information, we will deeply study the formation dynamics of abnormal aggregation of core proteins such as α-synuclein, TDP-43 and HTT in Parkinson's disease, ALS, HD, SCA and other diseases, the transmission pathways inside and outside cells, and their effects on neuronal function. The DIKWP model was used to analyze how the data perception, information transmission and processing processes were disturbed in the process of pathological aggregation, and how the purpose (such as the self-stabilizing regulatory mechanism) at the cell/individual level was out of control.

2. Constructing Innovative Disease Models and Simulating Cognitive Control Deterioration: Based on the theory of artificial consciousness and the hierarchical control framework of DIKWP, we will build advanced models that can simulate disease progression, including 3D brain organoid models derived from human induced pluripotent stem cells (iPSCs), and novel mammalian models that introduce the concept of DIKWP. Through these models, we will explore the relationship mechanism between "cognitive control layer disintegration" and neural network degeneration during disease development, including: how higher cognitive/control functions are gradually impaired when intracerebral protein pathology accumulates; Conversely, how does the loss of advanced cognitive control accelerate the deterioration of local pathology? Verify the effectiveness of the DIKWP framework in explaining the above interaction mechanisms.

3. Screening of disease intervention targets and design regulatory strategies: Based on the above-mentioned multi-level mechanism research, combined with systems biology and artificial intelligence methods, key molecules or network nodes (new drug targets) that can be used for intervention are identified. Intervention strategies for these targets (including small molecule drugs, gene therapy, immunotherapy, etc.) were designed, and the "DIKWP-drug regulatory pathway mapping model" was constructed using the DIKWP framework to simulate the multi-level impact chain of interventions on the disease system. The model predicts how potential therapies act on the data layer (molecular pathology) and progressively influences information processing, knowledge/network reorganization, Wisdom decision-making, and Purpose layer (functional improvement), guiding subsequent in vitro and in vivo experiments to validate the most promising treatment strategies.

Research content and technical route

Focusing on the above objectives, the project is divided into three interrelated research modules, which combine experimental research and theoretical modeling to advance simultaneously. The following diagram summarizes the overall technical roadmap of the project (from pathogenic mechanism analysis to model construction to intervention strategy), and the specific contents of each module are as follows:

Content 1: Analysis of pathological protein aggregation-propagation mechanism and DIKWP information dissonance

Main research question: How do pathological proteins such as α-synuclein, TDP-43, and HTT form abnormal aggregations in cells and spread in the nervous system through what pathways? How can the normal processing of information by cells and neural networks in these processes be disrupted, leading to dysfunction? We will study these questions using in vitro and in vivo model integration, and apply the DIKWP framework to unravel the mechanism of multi-level information dissonance.

Study Protocol:

• Multi-model experimental system: Build a series of models covering molecules, cells, tissues, and animals

  • In vitro molecular and cellular models: Stable transfection cell lines expressing mutant or tagged pathological proteins (such as human mutant α-synuclein, HTT-Exon1-Q_expanded, mutant SOD1 or TDP-43, etc.); In vitro recombinant protein aggregation models were used to analyze oligomer/filament formation kinetics; Co-culture of primary neurons is used to observe the transfer of pathological proteins between cells (e.g., seed-induced diffusion assays with labeled fibers).

  • Brain-like organoid models: Disease-associated brain region organoids were constructed from patient-derived iPSCs (e.g., midbrain substantia nigra dopaminergic organoids for PD, cortico-muscle motor neural circuit organoids for ALS, striatal organoids for HD, and cerebellar organoids for SCA). These organoids will be induced to form pathological protein accumulation to mimic early disease states.

  • Animal models: Transgenic mice (e.g., model mice overexpressing mutations α-syn, HTT) and novel model organisms (e.g., gene-edited rats, considering non-human primates if necessary) were used to validate in vivo. Viral vectors are used to inject pre-formed fibers of pathological proteins into specific nuclei of the mouse brain, and the pattern and speed of their transsynaptic transmission are tracked. Through in vivo imaging and tissue immunostaining, the propagation path of protein pathology in the brain was mapped.

• Pathological mechanism analysis: In the above model, the following indicators were determined and compared:

  • Aggregation kinetic parameters: threshold concentration, nucleation and elongation rates of different protein aggregation initiation, oligomeric structural characteristics, etc.

  • Cellular responses: stress responses triggered by the accumulation of abnormal proteins (e.g., activation or depletion of the autophagolysosomal system, dysregulation of the protein homeostasis network), key signaling pathways (e.g., inflammatory cascade, changes in oxidative stress levels).

  • Neuronal function: Neuronal membrane potential and firing mode changes, synaptic transmission function (postsynaptic current amplitude and frequency changes) and neural circuit activity changes were recorded (multi-electrode array MEA to detect organoid network synchronization, calcium imaging to observe neuronal network activity).

  • Transcellular Propagation: Quantify the extent to which a tagged protein spreads from the initial site to the distal site, assessing whether it spreads along anatomical junctions and whether there is a transsynaptic release-uptake mechanism. Compare the propagation characteristics of different proteins in different models to find common rules and differences.

  • Human correlation: Validate model findings by combining patient brain tissue samples and genetic data. For example, whether the α-syn aggregation region in the brain tissue of patients with SNCA duplications/mutations associated with Parkinson's disease is consistent with the model transmission pathway; Whether the distribution of TDP-43 pathology in the spinal cord of ALS patients conforms to the transmission sequence predicted by the model. Genome-wide association study (GWAS) data was used to focus on susceptibility genes related to protein homeostasis and propagation (such as autophagy-lysosomal pathway genes) and test their functions in the model.

• DIKWP Model Interpretation: Based on the above data, we will use the DIKWP framework to unravel the information processing dysregulation in the aggregation and propagation of abnormal proteins:

  • Data layer: The initial signal of pathological triggers is equivalent to the "data" perceived by the cell, such as the appearance of misfolded proteins, concentration exceeding threshold, etc. Normally, cellular quality control mechanisms feed this data into subsequent processes (e.g., ubiquitination degradation pathways). We analyze whether cells perceive these data in disease states for retardation or allergies, and for "misreadings" (e.g., misidentification of misfolded proteins as harmless).

  • Information layer: Cells process raw data to form "information" that corresponds to processes such as signaling pathways being activated and transcription factors responding. We examine aberrant protein-induced signaling network restructuring: which key pathways are activated (e.g., inflammatory cytokine signaling) and whether the direction of information flow deviates from homeostasis (e.g., persistent stress signals lead to apoptosis tendency). From the perspective of DIKWP, this step is equivalent to the process of removing noise and adding context, and we will identify the deviations in the information extraction process in the disease cells (e.g., the noise is amplified instead of being filtered out).

  • Knowledge layer: Information is further integrated and abstracted into "knowledge", which corresponds to the cells to establish cognition and coping strategies for environmental conditions, such as the formation of memories through feedback regulation of gene expression patterns. We will investigate whether chronic protein accumulation alters the cell's gene expression program (similar to "false learning"): for example, neurons may reduce synapse-associated protein synthesis to compensate, and stress persistence forms a new constant expression profile. The knowledge layer of the DIKWP model emphasizes the grasp of objective laws, and we will look for the "wrong knowledge" formed by disease cells (such as the persistence of abnormal proteins as the new normal), which leads to weakened defense mechanisms or misallocation of resources.

  • Wisdom Layer: Wisdom corresponds to system-level decisions and behaviors. For neural networks, this can be compared to the overall functional output and coordination. In this layer, we investigate whether the pathological accumulation of proteins causes changes in network properties (such as changes in network synchronization and structural connectivity recombination) at the loop level, resulting in a decrease in the system's decision-making ability. For example, in Parkinson's disease, the brain's ability to make "wisdom decisions" in motor control as a whole is reduced due to damage to the substantia-striatal circuitry (manifested by bradykinesia, erroneous movements). Computational modeling, such as neural network modeling, simulates basal ganglia dynamics, to validate how pathological burden alters network decision attributes. From the perspective of DIKWP, this is a manifestation of the hindrance or distortion of the transformation of knowledge into wisdom.

  • Purpose Layer: As the highest level, Purpose represents the goal, will, or motivation of the system. In organisms, it can correspond to the top-level regulatory purpose of maintaining homeostasis and achieving normal behavior. This project will explore the disintegration of the purpose level in the context of disease: the purpose of the normal brain (e.g., maintaining smooth movement, maintaining cognitive function) may be "out of control" or "offset" due to pathological invasion. For example, impulsive behavior and decreased volitional behavior in HD patients due to degeneration of the striatum-frontal circuit can be seen as impaired ability to drive purpose. We intend to capture the trajectories of the function of the Purpose layer during disease progression through animal behavior experiments and patient high-level functional assessment (such as executive function tests), and analyze its relationship with the degradation of the lower level Wisdom/Knowledge layer. From the perspective of DIKWP, a "Purpose Runaway Model" was established to explain how the activities of the lower circuit become disordered when the top-level purpose cannot be constrained normally, which further aggravates the pathological diffusion.

Expected results: Through the above research, the first content will generate a new understanding of the pathogenic role of core pathological proteins, including their aggregation and propagation laws and the resulting signal network reconstruction. More importantly, we will obtain a multi-level analysis of the lesion process under the unified framework of DIKWP: from the initial molecular data abnormality, to the cell information processing disorder, to the neural circuit Wisdom decision-making impairment, and the whole-process mechanism model of the disintegration of the target layer control. This will provide a rationale for subsequent interventions targeting high-level functions.

Content 2: Construction of disease progression model based on DIKWP and research on the mechanism of "cognitive control layer disintegration".

Main research question: How to construct a model that can reproduce the whole process of neurodegenerative diseases, especially the deterioration of higher cognitive/control functions? How does the loss of high-level cognitive control interact with the structural degeneration of the underlying neural network during disease progression?

Study Protocol:

• Establishment and improvement of brain organoid disease models: The human brain organoid models established in content 1 (containing familial mutations or induced pathological protein accumulation) were used to further culture and induce the emergence of advanced disease characteristics. This includes extending the culture time to promote age-related changes in organoids (e.g., protein accumulation, neuronal subtype loss, glial cell activation), and adding external stimuli/stress (e.g., α-syn preformed fibrils to stimulate organoids, glutamate excitotoxicity stimulation, etc.) to accelerate neurodegeneration. Due to the lack of sensory input and environmental interaction of conventional brain organoids, it is difficult to generate high-level neural functions, and this project will try to enhance the function of organoids by combining artificial consciousness theory. Specific methods include:

  • The organoids are connected to the microelectrode array (MEA) dish, and their electrical activity is recorded in real time and electrical stimulation feedback is given, so that the organoids form a closed-loop interaction with the outside world. For example, referring to the "DishBrain" study, an algorithm was designed to give reward/punishment stimuli based on the organoid firing pattern, so that it "learned" to produce a specific firing response. Through this training, certain goal-directed behaviors are introduced to the organoids (e.g., learning simple pulse rhythms, synchronized with external signals).

  • Under the guidance of the DIKWP framework, organoids are given a "Purpose Layer" simulation: a high-level decision-making algorithm (based on reinforcement learning or DIKWP cognitive architecture) is run on the computer as a "virtual Purpose Layer" of organoids. The algorithm uses the electrophysiological data of organoids as input (data/information) and outputs guidance signals to act on organoids (affecting organoid activity through optogenetics or electrical stimulation) to simulate the regulation of the lower layers of the brain from the upper levels. In this way, an artificial cognitive control loop is constructed to realize an organoid-computer hybrid system. We will explore the differences in the behavior of the system in healthy and disease states: whether the healthy organoids + control system can better maintain stable activities, and whether the diseased organoids + control system have control signal disorders and coordination failures.

  • Organoids were observed to see if there were more complex patterns of activity, such as periodic rhythms, electrical band segmentation (e.g., EEG-like δ/thetama/β waves), and simple learning-memory representations (e.g., stimulus-response associations). It focuses on the evaluation of indicators related to cognitive control, such as the learning index of organoids in response to external changes, the ability to generate autonomous rhythms, etc. Monitoring the decay trajectories of these higher functions over time in disease organoids as an in vitro model of "cognitive control layer disintegration".

• Development of novel mammalian models: Although rodents have been widely used in the study of neurodegenerative diseases, their higher cognitive functions are limited, and in order to simulate the "disintegration of the cognitive control layer", this project considers the introduction of models closer to humans or specially designed animal models

  • Non-human primate models: When necessary, work with the primate disease model research team to utilize transgenic monkey models (e.g., macaques carrying mutant α-syn or HTT). Non-human primates have more developed prefrontal lobes and cognitive functions, which can assess higher functional impairments such as cognitive impairment in Parkinson's disease and mental and behavioral symptoms of HD, and more closely resemble the degenerative manifestations of the human "cognitive control layer".

  • Design of a mouse model with controllable cognitive modules: Inspired by artificial consciousness, an attempt was made to introduce an "artificial control layer" through optogenetics in transgenic mice. For example, light-sensitive channels are expressed in a specific area of the mouse brain (prefrontal cortex), and an external computer is connected to monitor their motor/cognitive behavior in real time, and the light stimulation is regulated according to the DIKWP algorithm. This attempt was made to give mice an external cognitive control circuit to explore how symptoms of pre-existing disease (e.g., motor disorders, cognitive task performance) changed when the circuit was impaired (by stopping the stimulation or by incorrectly stimulating). Although challenging to implement, this exploration could provide evidence for validating the impact of "control layer disintegration" on disease.

  • Behavioral Assessment: Indicator Extraction of "Cognitive Control Layer Disintegration": In animal models, a series of behavioral tasks are introduced to evaluate advanced cognitive control functions, such as maze learning (spatial cognition), flipped learning tasks (cognitive flexibility), Go/No-go or stop signaling tasks (inhibitory control), etc. These tasks target executive function and cognitive control, and Parkinson's disease model mice (dopamine deficiency) and HD model mice (striatal degeneration) are expected to exhibit significant deficits in these tasks. The success rate of the above tasks and reaction time were recorded in the course of the disease, which were used as quantitative indicators of cognitive control ability with the development of the disease. Combined with neural recording technology (such as implanting microelectrode arrays in the brains of free-moving animals to monitor cortex-striatal-thalamic circuit discharge), we can analyze how the neural network activity pattern changes with the deterioration of the disease during the execution of cognitive tasks, and identify the early signals of network disintegration.

• Analysis of the interaction mechanism of "cognitive control layer disintegration": Based on the results of organoids and animal models, the causal relationship between the deterioration of advanced cognitive control function and underlying neurodegeneration is analyzed from the perspective of DIKWP

  • When the underlying neurons are degenerated and protein pathologies accumulate, how do the higher-level purpose signals (such as brain instructions) change? We will test the hypothesis that low-level faults can be propagated upwards to disrupt higher-level control. For example, by comparing disease models whether the prefrontal cortex (associated with cognitive control) has neurodynamic changes early in the lesion (even if the pathology is predominantly in the underlying motor area). If it is found, it means that the low-level pathology spreads to the high-level pathology, resulting in the gradual disintegration of the control layer.

  • Conversely, when high-level control is impaired (e.g., by inhibiting prefrontal activity by human intervention in the model), does it exacerbate low-level pathology? For example, in Parkinson's disease model mice, frontal lobe executive function training (e.g., cognitive training) may have an ameliorating effect on motor symptoms; Do motor symptoms worsen more quickly if they are deprived of cognitive stimulation? Positive cognitive activity has been shown to slow neurodegeneration. We will verify the protective effect of high-level activity on the survival and function of lower neurons by controlling the level of cognitive stimulation in animals, organoid stimulation patterns, and other experiments. If true, it would suggest that cognitive interventions can also be used as therapeutic ideas.

  • The DIKWP model is used to conceptualize the above two-way relationship: the high-level cognitive control is regarded as the top-level "Purpose/Wisdom" module, and the underlying neurodegeneration is regarded as the "data/information" module failure. We will try to build a simplified mathematical model that represents the interaction between the two in terms of differential equations or cybernetic models (e.g., Wisdom layer output = f (knowledge layer state, Purpose signal), knowledge layer decay rate = g (pathological load, Wisdom layer input)). By fitting experimental data, we look for the model form that best explains the cognitive-pathological interaction, and refine the key parameters (representing the intensity of the "gain" or "gating" effect of a certain biological process, such as the Purpose layer). This model will provide a quantitative tool for the evaluation of intervention strategies for both the top and the bottom in Content 3.

Expected Results: Content 2 will produce innovative disease models and new understanding of the mechanisms of cognitive control deterioration: for the first time, brain-like organoids and animal models with artificial cognitive control circuits will be constructed to simulate the impairment of higher functions in neurodegenerative diseases. On this basis, we will elucidate the causal loop between cognitive control layer disintegration and neural network degeneration, and answer the core questions of how the decline of high-level functions accelerates the degeneration of the underlying layer and how the underlying lesions affect the high-level functions. The results of this part will provide theoretical support for disease intervention from a cognitive perspective, and also provide a more comprehensive model platform for subsequent evaluation of new drugs/therapies.

Content 3: Screening of potential new drug targets, design of intervention strategies and simulation of DIKWP-drug effect pathways

Main research questions: Based on the above-mentioned mechanistic studies, identify the key nodes (molecular, cellular, or circuit level) that can be targeted for intervention in the disease process, and develop corresponding therapeutic strategies. How can we assess the impact of an intervention at all levels of the disease system, from molecular to cognitive? Can the DIKWP framework be used to construct a multi-level transmission model of drug action to assist in predicting efficacy and optimizing the combination of strategies?

Study Protocol:

• Target Mining and Validation: Based on the data obtained in content 1 and 2, systems biology and bioinformatics methods were used to screen potential intervention targets

  • At the molecular/cellular level, differential omics analysis (transcriptome, proteome, metabolome) is performed to look for molecules that are significantly dysregulated in disease states. For example, which upstream regulators in the cellular stress network induced by protein aggregation change the most (potential master regulatory switch); Which metabolic pathways are most severely impaired during neurodegeneration (e.g., key enzymes for mitochondrial function). Combined with the database of known drug action targets, match whether there are molecules that can be regulated by drugs. Priority is given to key nodes that are abnormal in multiple disease models as candidate intervention targets.

  • At the neural circuit level, according to the results of content 2, the nodes that play a key role in maintaining the stability of the network are identified. For example, if the intensity of activity in a particular brain region (eg, prefrontal cortex, specific nuclei of the basal ganglia) is highly correlated with the severity of disease symptoms, this brain region/circuit may be a breakthrough for intervention. If possible, chemogenetics/optogenetics were used to verify whether stimulation or inhibition of this node would improve model animal performance.

  • Using genetic clues: Finding out the pathways involved in the disease risk gene set (such as the lysosome/mitochondrial pathway of PD, the RNA metabolism pathway of ALS, etc.) from the large-scale gene association data of humans, and the corresponding pathway molecules can be used as potential targets. For example, the frequent mutations of LRRK2 and GBA genes in PD suggest that their pathways are worth targeting. C9ORF72 pathological mechanisms in ALS involve ribosomal stress and may also be considered. Starting from the cross-disease commonality, pay attention to whether there are multiple susceptible pathways shared by diseases, and if so, their key nodes are more valuable for general intervention.

  • The above list of candidate targets was optimized through multi-dimensional weight scores (such as validation results in several models, drug accessibility, whether it affects high-level functions, etc.), and finally some of the most promising new targets were determined.

• Intervention strategy design and preliminary evaluation: Formulate corresponding intervention plans around each key target selected:
  For each candidate intervention strategy, a preliminary evaluation is carried out in the cell/organoid model to detect whether it can reduce pathological protein load, restore cell viability, and improve the synchronization of organoid electrical activity. If a strategy shows good results in vitro, it is advanced to animal models to test its behavioral and pathologic improvements (e.g., whether drug administration delays motor function deterioration or improves cognitive task performance in model mice).

  • Small molecule drugs: Screening existing or designing new small molecule inhibitors/activators. With computer-aided drug design (CADD) and virtual screening, find candidate molecules from a library of compounds that bind to key sites of the target protein. For existing marketed drugs, consider the possibility of repurposing for indications. For example, some anticancer or metabolic drugs act on shared pathways and may interfere with neurodegenerative processes. There have been successful cases in this area, such as metformin (diabetes drug) that has been studied to improve neuronal metabolism to combat AD/PD pathology.

  • Biotechnic therapeutics: designing gene therapies (e.g., silencing disease-causing genes using AAV vectors to overexpress protective genes or RNA interference), antibody therapies (e.g., monoclonal antibodies to remove extracellular aggregates), and cell therapies (e.g., transplanting genetically modified neural stem cells to secrete beneficial factors). Particular attention is paid to whether high-level cognitive control can be enhanced to intervene in the disease, such as electrical stimulation/brain-computer interface assistance.

  • Combination strategy: Consider the regimen of multi-target combination intervention, because dyskinetic degenerative diseases often have complex mechanisms, and multi-level combination intervention may be more effective. The DIKWP model helps to pick and choose combinations that act on different levels. For example, one regimen may include a two-pronged approach of "molecular layer clearance protein aggregation + loop layer brain stimulation" to simultaneously alleviate underlying pathology and apical dysfunction.

• Construction of DIKWP-drug regulatory pathway mapping model: After obtaining some experimental validation, this project will develop a computer simulation platform to combine the DIKWP framework with the drug action model to simulate the global effect of predicting the intervention strategy. Here's how:

  • Model framework: Based on the five-layer structure of DIKWP, model nodes corresponding to biological levels are established: the data layer corresponds to key molecular/pathological indicators, the information layer corresponds to the state of cell signaling pathways, the knowledge layer corresponds to the local network structure/plasticity, the Wisdom layer corresponds to the overall network function indicators (such as network efficiency and synchronicity), and the Purpose layer corresponds to the behavior output or cognitive score. The causal relationship between the layers will be depicted in the form of mathematical relationships on the basis of the above experiments and literature, for example, the increase of pathological protein load (data) leads to the activation of apoptosis signals (information), which can be modeled as the activation function of data to the information layer; For another example, the decline of network function (Wisdom) will affect the realization of the Purpose, which is expressed as a function of the output of the Wisdom layer that determines the achievement of the goal of the Purpose layer.

  • Drug effect embedding: For each intervention, a corresponding effect is added to the model: for example, small molecule drug X can reduce the variable of "aggregation level of a protein in the data layer" by a certain proportion, gene therapy Y can enhance the "information layer autophagy pathway activity" and indirectly reduce the pathological production rate of the data layer, and brain stimulation Z can improve the "Wisdom layer network synchronization index", etc. Through the literature and our experimental data, the range of quantitative parameters for these effects is given.

  • Simulation and optimization: Numerical simulation of different intervention programs was carried out using the constructed model. After the intervention was introduced, the changes in the key indicators of each layer of the model over time, as well as the improvement of the final output of the "Purpose layer" (which can be understood as symptom score or functional score). For example, comparing the simulated effects of drug X alone, brain stimulation Z alone, and X+Z combined to find the combination that had the greatest improvement in the top-level Purpose indicator. At the same time, the model can be used for sensitivity analysis: to assess the impact of the uncertainty of each parameter on the outcome, so as to understand the possible changes in the treatment strategy under different individuals (different parameters). This is akin to building a digital twin to test therapies in a virtual environment and compare the pros and cons.

  • Feedback optimizes experimental design: The simulation results will be used to guide subsequent real-world experiments. For example, if the model predicts that the combination therapy X+Y synergy is significantly superior to that of a single agent, we will focus on validating this combination in an animal model; If the model suggests that a key parameter (e.g., drug dose-effect curve) has a significant impact on efficacy, we will design experiments to measure this parameter in detail and calibrate the model. Through the iteration of models and experiments, the understanding and success rate of disease intervention are improved.

Expected Results: Content 3 will eventually produce several promising intervention targets and therapeutic strategies, and verify their effectiveness through preliminary experiments; More importantly, we will get a "digital model of disease-drug action" that integrates the DIKWP theory. This will reveal a panoramic picture of how the intervention works at all levels, helping to uncover indirect pathways or potential risks that traditional approaches fail to detect. For example, a model may show that a drug targeting a molecular target reduces protein aggregation but causes a compensatory change in the knowledge layer that has little effect on the Wisdom layer, alerting the need for other means. The research in this content will provide a basis for the optimization of strategies before clinical trials enter human trials, and realize an important leap from mechanism discovery to application translation.

Feasibility analysis

The project is novel and ambitious, but technically and theoretically feasible, mainly reflected in:

1. Preliminary Research Basis: The project team has carried out extensive preliminary work in related fields. In terms of theory, Professor Duan Yucong's team, as a leader in the field of DIKWP model and artificial consciousness, has published a series of results and has a complete independent intellectual property system. It includes proposing the basic principles of the DIKWP model, constructing a "dual circulation" artificial consciousness architecture, and developing the DIKWP semantic operating system. In particular, the team has established the DIKWP Artificial Consciousness Laboratory and developed the DIKWP-AC Physiological Artificial Consciousness Prototype System, which has initially realized the mapping of the DIKWP model to the neurophysiological process of the brain. This prototype system was presented at the first World Congress of Artificial Consciousness and demonstrated the feasibility of using the DIKWP model to simulate cognitive processes in the brain. These works provide a solid theoretical and instrumental foundation for this project, which enables us to apply DIKWP to biomedical exploration. In terms of biomedicine, many core members of this project have been engaged in the research of the mechanism of neurodegenerative diseases for a long time, and have accumulated rich experimental experience and resources. The team has mastered the breeding and phenotypic analysis technology of mouse models of Parkinson's disease, ALS and other diseases, established an iPSC cell line library and successfully differentiated brain organoids, and published papers on protein aggregation and neuroinflammation mechanisms in international journals. We have the technical platform and talents to carry out key experiments such as molecular biology (such as viral vector construction, CRISPR gene editing), omics analysis (transcriptome sequencing, proteome quantification), electrophysiological recording (patch-clamp, MEA), etc. The team also worked with multiple hospitals to obtain patient clinical samples and data that could be used to validate the clinical relevance of the findings. These upfront foundations ensure that the project is "thought of and done".

2. Feasibility of technical route: The technical means to be used in the project are all mature or rapidly developing methods, and have been applied in our team or cooperative units

• Protein aggregation and propagation studies: Related cell models (e.g., α-syn fiber-induced diffusion) and animal models (intracerebral injection of pathogenic proteins) have been well reported in the literature. In recent years, our laboratory has also carried out research on α-syn propagation and accumulated the necessary technology. Pathological section immunoassay and fluorescence quantitative imaging can accurately track the diffusion path of aggregates.

• Brain organoids and disease modeling: We have successfully cultured region-specific organoids such as cortex, hippocampus, and substantia nigra and observed spontaneous electrical activity. Organoids have been used internationally to study the pathology of Alzheimer's disease and Parkinson's disease. The introduction of external electrical stimulation and closed-loop control into organoids is a new attempt, but the required components (multi-electrode arrays, stimulus-feedback algorithms) are already commercially available and can be integrated by our team's engineering staff.

• Artificial Consciousness System Integration: The DIKWP-AC prototype system is already in its infancy. Docking it with organoids/animal models requires custom hardware and software interfaces, but we have a basis for collaboration in the field of brain-computer interfaces. In addition, there have been similar successful cases of reinforcement learning algorithms used in biofeedback training, and the required software framework can be quickly developed by drawing on open source solutions.

• Omics and computational modeling: The multi-omics sequencing and analysis platform is well-equipped in our team and partner units, and the data processing pipeline is mature. The DIKWP-drug pathway model requires some software development and mathematical modeling, but the team has members who are familiar with systems biology and computational modeling, and can be supported and guided by data from Prof. Duan's DIKWP laboratory. Previous literature has explored the introduction of control theory into disease mechanism and drug development, which proves that this idea is feasible, and our DIKWP model will further enhance the explanatory power.

3. Project management and cooperation support: This project is supported by the school and multidisciplinary platform of the research group, and has a good implementation guarantee. The project team has formulated a detailed implementation plan and risk plan, and each module is smoothly connected and can be carried out relatively independently to ensure flexible adjustment. The funds are mainly used for the purchase of reagent consumables, organization of laboratory animals and data analysis and computing resources, and the budget arrangement is reasonable and feasible. We have established cooperation with top research institutions and hospitals in China, such as in primate model culture, patient sample acquisition, etc. Once a grant has been awarded, critical experiments and model building can begin immediately. The project leader and key researchers have experience in hosting national projects, and are familiar with the organization and management of large-scale projects, which will ensure that the project is carried out in an orderly manner as planned.

To sum up, this project has a good foundation in terms of personnel reserves, experimental conditions and theoretical tools. The proposed methods are bold and innovative but well-founded, and we will implement each step with a solid experimental and scientific attitude. The multidisciplinary nature of the project also ensures that in the process of implementation, even if a certain technical route encounters challenges, the team can quickly adjust the plan, with a high risk resistance and success probability.

Innovation

This project combines cutting-edge theories of artificial intelligence with neuroscience challenges, and is expected to produce the following innovations:

1. Pioneering Theory Fusion: The DIKWP model and artificial consciousness theory proposed by Professor Duan Yucong were applied to the research of neurodegenerative diseases for the first time. The five-level cognitive framework is used to analyze the biopathological process, and the perspective of "Purpose" is introduced in the field of neurodegeneration mechanism research, creating a new research paradigm. This cross-border convergence is expected to break through the limitations of traditional neurobiology that only looks at problems at the molecular or cellular level, and provide a holistic understanding of disease.

2. Brain-like organoid-artificial consciousness hybrid model: A new disease model is proposed and practiced, which connects human brain organoids with artificial intelligence control systems, and endows organoids with certain environmental interaction and goal-oriented capabilities. We will simulate the "collapse of cognitive control" in diseases, a phenomenon that was difficult to reproduce in the laboratory before. The establishment of this model is the first of its kind in the world, opening up a new way for disease mechanism research and drug screening.

3. Integrated analysis of multi-level pathological mechanisms: Using the DIKWP framework as a guide, the disease mechanism is studied from multiple levels such as molecular aggregation dynamics, cell signaling networks, neural circuit functions, and cognitive behavior, and the causal relationship between each level is clarified. In particular, the concepts of "information processing dysregulation" and "purpose out of control" are proposed to explain how protein aggregation leads to dysfunction. This cross-scale integrated analytical approach is unprecedented in the study of movement disorders, and is expected to uncover key mechanisms that cannot be revealed by traditional single-level research.

4. DIKWP Drug Effect Pathway Model: Innovatively construct a DIKWP-drug regulation pathway mapping model to simulate the impact chain of interventions on the disease system. Unlike traditional pharmacodynamic models, which focus only on biochemical effects, our models cover the entire chain from molecular to cognitive and are able to evaluate the impact of interventions on higher order functions. This model can be used to optimize the design of multi-target combination therapies and improve the success rate. This will be a new drug discovery aid that can be scaled up to the design of therapies for other complex diseases.

5. Potential common targets and strategies: By comparing the similarities and differences of Parkinson's disease, ALS, HD, SCA and other diseases, we are expected to propose cross-disease generic intervention targets (such as protein homeostasis maintenance pathways, neuroinflammation regulators) and global prevention and treatment strategies (such as cognitive stimulation therapy). This exploration of the idea of "treating different diseases at the same time" for movement disorders will enrich the treatment theory of neurodegenerative diseases and have the potential to bring a broad spectrum of new therapies.

In summary, the innovation of this project lies in the theory, methodology, tools and strategies. Its implementation will inject the latest concepts of artificial intelligence into the research of neurodegenerative diseases, accelerate the transformation of theory and practice in this field, and is expected to achieve original results with international influence.

Research Basis

The application team of this project has a deep research foundation and good cooperation conditions in related fields, and can provide strong support for the smooth implementation of the project:

1. Academic foundation of the project leader and main members: The project leader, Professor Duan Yucong, is a well-known expert in the field of artificial intelligence and cognitive computing, and has made outstanding achievements in the DIKWP artificial consciousness model. As the initiator of this model, Professor Duan has published dozens of related papers and obtained more than 100 domestic and foreign invention patents. These achievements have established his position in the international academic community of artificial consciousness. Professor Duan has an interdisciplinary background, and in recent years, he has actively combined artificial intelligence theory with life sciences, and has guided the development of the DIKWP physiological and artificial consciousness prototype system and tried to apply it to medical scenarios. Another core member, Dr. Tang ××, presented the work of the DIKWP white-box evaluation framework at the 2023 international conference, laying the foundation for the interpretability of the DIKWP model. The above work shows that the team has outstanding strength in theoretical innovation and software development.

In terms of biomedicine, the project team has also brought together outstanding talents in neuroscience and bioinformatics: associate PI researcher Li ×× has been engaged in the study of the mechanism of Parkinson's disease for a long time, discovered new evidence of α-syn transmission, and published many papers in journals such as Acta Neuropathologica; Prof. Wang ××, Associate PI, is an expert in genetic and neurological diseases, who has conducted in-depth research on the pathogenic genes of ALS and SCA, and has presided over relevant projects of the National Key R&D Program. Among the young backbones of the team, Dr. Zhang × specializes in stem cell and organoid technology, and has published research papers on brain organoids; Dr. × Liu is proficient in multi-omics data analysis and computational modeling. In their respective fields, they have a series of high-level results, forming a solid support for the objectives of this project.

2. Previous work achievements: The previous relevant research of the project team members provides preliminary data and empirical support for this topic

• In terms of protein aggregation and cellular response, we have observed that abnormal α-syn aggregation can lead to autophagic lysosomal dysfunction and dopaminergic neuron damage using Parkinson's model cells, and the results have been published in a paper. This proves that we have mastered the methods for protein aggregation detection and cellular stress analysis.

• In terms of animal models, the team successfully established transgenic α-syn mice and mutant HTT (R6/2) mice, and conducted behavioral follow-up to obtain data curves for the deterioration of motor and cognitive function with age in these models. This provides a baseline for subsequent evaluation of the effect of the intervention. We also attempted to inject pre-α-syn fibers into ordinary mice, which resulted in Lewy-like pathology at synaptic junctions, confirming the prion transmission hypothesis.

• In terms of brain organoids, we have successfully verified the formation of HTT aggregates in organoids by culturing 3D brain tissue with a diameter of 4 mm containing cortex and striatum regions, in which stem cells from HD patients have been introduced. This previous experiment showed that organoids could partially reproduce the developmental abnormalities and neurotoxicity of Huntington's disease.

• In terms of artificial intelligence simulation, we have conducted biological verification experiments of the DIKWP model internally: we used the developed software to simulate the degradation process of the neural network, tried to reconstruct the phenomenon of "cognitive control layer disintegration" in the model, and compared it with the degradation model of the human brain functional network reported in the literature, and obtained a preliminary agreement. This gives us the confidence to build more complex digital twins of DIKWP diseases.

• In terms of drug screening, we have conducted drug retargeting studies for Parkinson's disease using gene expression data and drug gene signature banks to identify several possible drug candidates (e.g., drugs that regulate lysosomal function). Some drugs have been tested at the cellular level and have been shown to have an inhibitory effect on α-syn aggregation. This experience can be directly used in the screening process of content 3 of this project.

3. Platform and cooperation: The unit has advanced scientific research platforms, including a molecular biology center, an animal experiment center (including an SPF-level animal room and a behavioral laboratory), a cell and organoid culture room, and a high-performance computing center. These platforms can meet the experimental and analytical needs of the project in all aspects. In addition, we have established synergistic relationships with many teams at home and abroad, such as cooperating with a top AI research institute in China to develop DIKWP model applications, sharing organoid culture experience with a foreign neuroscience laboratory, and cooperating with the neurology department of an affiliated hospital to collect clinical samples and carry out translational research. These collaborations will provide a strong impetus for project implementation and provide a channel for the output and replication of results.

Based on the above research foundation, we believe that the project team has sufficient strength to achieve the set goals. No matter from the perspective of personnel knowledge structure, existing achievements accumulation or experimental conditions, it provides a guarantee for the smooth development of this project. The team's past experience in working together will also ensure close cooperation and resource sharing between various tasks. It can be expected that during the implementation of the project, we will continue to produce high-quality medium-term results to support the achievement of the final overall goal.

Team members and schedules

Project Team Composition: This project is undertaken by a multidisciplinary team, and the core members include:

• Yucong Duan (Project Leader): Professor, expert in artificial intelligence and cognitive computing. Responsible for the overall idea planning, DIKWP theoretical guidance and artificial consciousness system development. Coordinate the collaboration of each research group and be responsible for key decisions and milestones of the project.

• Li ×× (Associate PI): Researcher, neurobiology expert. He is responsible for mechanistic experiments in Parkinson's disease and ALS, including cell/animal model construction and molecular mechanism analysis. Assist in the application of the interface DIKWP model to these mechanisms.

• Wang ×× (Associate PI): Professor, expert in genetics and neurology. He is responsible for HD and SCA research, as well as the application of human genetic data in the project. Coordinate the analysis of patient samples and guide the mining of genetic targets.

• Zhang × (Young PI): Associate Researcher, Stem Cell and Organoid Technologist. Responsible for the cultivation of brain-like organoid models and disease simulation experiments, and the realization of the interface with artificial consciousness.

• Liu × (Young PI): Associate Researcher, Expert in Bioinformatics and Systems Biology. Responsible for multi-omics data analysis, DIKWP-drug pathway model construction and computational simulation.

• Other members: Several postdoctoral fellows and graduate students are involved in the experimental execution, data collection and analysis of each module. In addition, two clinical consultants (the director of the neurology department of a hospital, etc.) assist in the acquisition and interpretation of clinical samples and data.

The team structure covers theory, experiment, and calculation, and the personnel have a clear division of labor and mutual cooperation to ensure the smooth progress of the multi-dimensional tasks of the project.

Planned schedule: The proposed implementation cycle of the project is 5 years, and the research content will be promoted in stages

• Year 1: Model Preparation and Basic Experimentation Phase.

  • Content 1: Establish a pathological protein cell model and a first-generation neuronal co-culture system; In vitro aggregation experiments such as α-syn and TDP-43 were carried out to optimize the detection methods. Preliminary observation of the stress response of cells to protein aggregation (omics pre-experiment).

  • Content 2 initiation: The first batch of disease-related brain organoids (such as organoids containing mutant HTT) were cultured to confirm their morphological and preliminary pathological characteristics. The MEA recording system was built and integrated with organoid culture.

  • Content 3 Initiation: Collect and sort out multi-omics and literature data, and preliminarily screen disease-related pathways and potential drug targets. Establish the basic computing framework of the DIKWP model.

• Year 2: Mechanism in-depth research and model refinement.

  • Content 1: Expand to animal models to monitor pathological protein diffusion trajectories (immunohistochemistry, in vivo imaging) in transgenic mice; Perform key signaling pathway analysis (transcriptome sequencing to identify aggregation-induced pathway changes); The experimental data were mapped to the DIKWP framework layer by layer to form a preliminary multi-layer mechanism model.

  • Content 2: Introduction of artificial control: Develop a closed-loop stimulation algorithm for organoids, and start feedback training on organoids to evaluate the increase in the complexity of their electrical activities. A cognitive behavior testing platform for Parkinson's model mice was established to obtain the baseline difference between normal and model mice on cognitive tasks.

  • Content 3: Based on the data of the first year, a number of targets were identified and the drug/gene intervention experimental screening at the cellular level was carried out. For example, the effect of lysosomal enhancers on α-syn accumulation is tested. The structure of the DIKWP drug model was optimized, and the parameters were calibrated based on the preliminary intervention data.

• Year 3: Validation & Synthesis Phase.

  • Content 1: Design a verification experiment for the mechanism hypothesis proposed by the DIKWP model analysis. For example, if the model indicates that an information pathway is overactivated in a disease, drug intervention to see if neuronal damage can be mitigated. Integrate the results of multiple diseases and refine the common mechanism.

  • Content 2: Complete the construction of a mixed model of organoids + artificial consciousness, and compare the performance differences between healthy and disease organoids in cognitive control tasks. For mice, optogenetic modulation of the prefrontal lobe was attempted to verify the effect of cognitive activity on symptoms. A large amount of data was acquired to construct a cognitive-pathological interaction model.

  • Content 3: Advance 1~2 candidate intervention strategies confirmed to be effective to mouse model testing to evaluate their improvement effect on motor and cognitive phenotypic and pathological indicators. Improve the DIKWP-drug effect model so that it can reproduce the effects of these interventions and provide a basis for further optimization.

• Year 4: Optimization & Improvement Phase.

  • Content 1: Supplement mechanistic issues that have not yet been fully elucidated, such as the behavior of specific proteins in special environments (e.g., aggregation properties at different pH and metal ion conditions). Refine the details of the DIKWP mechanistic model to make it applicable to all target diseases.

  • Content 2: Based on the results of the previous year, the hybrid model was improved (e.g., adding more sensory input dimensions and improving the maturity of organoids) to further observe the detailed process of cognitive control disintegration. Write a paper on the model methodology and publish and share the innovative model.

  • Content 3: Synthesize the simulation results of the DIKWP drug model, optimize and adjust the tested intervention strategies (such as dose combination optimization), and explore new combination regimens. Start writing a paper on the effectiveness of the intervention.

• Year 5: Integration & Closing Phase.

  • Content 123 cross-integration: Hold a special seminar to cross-validate mechanism discovery, model research and intervention strategies. For example, the use of intervention strategies to validate the importance of a certain mechanism, or conversely, to predict a new combination of therapies that are more effective.

  • Finally, a complete map of disease mechanisms and intervention strategies under DIKWP was formed, and a comprehensive paper was written. Preparation of relevant patent applications (e.g. DIKWP drug evaluation software).

  • At the same time, a follow-up research plan will be formulated, such as obtaining effective data and models to apply for preclinical research projects, so as to prepare for real clinical translation.

In the implementation of the project, we will hold an internal progress meeting every six months, and invite consultants and experts to evaluate and guide every year, find problems in time and adjust the direction. Milestones are set up for important milestones (such as model establishment and key data output) to ensure that they are completed on schedule. Team management will focus on multidisciplinary communication, give full play to the expertise of each member, and achieve a synergistic effect of 1+1>2.

In short, the project has a clear timetable and a reasonable allocation of tasks. Within the established 5 years, we are confident that we will complete the research content with high quality and achieve the expected goals. Through the implementation of this project, it will not only deepen the scientific understanding of the mechanism of neurodegenerative diseases such as movement disorders, produce a series of high-level papers, but also give birth to new disease models and drug evaluation tools, lay the foundation for future research and drug development, and cultivate interdisciplinary scientific research talents. We believe that this exploration of integrating artificial intelligence and life sciences will open up a new direction for overcoming difficult diseases such as Parkinson's disease, and has important academic value and potential social benefits.