Upfront Basics

The neural mechanisms of sleep and wakefulness regulation and the diseases caused by disorders are one of the important frontiers of neuroscience and clinical medicine. Sleep disorders, including insomnia and circadian rhythm sleep disorders, are highly prevalent in the population, with about 10–35% of the general population reporting long-term insomnia. Sleep deprivation or rhythm disruptions not only affect memory, attention, and metabolic function, but also increase the risk of cardiovascular disease, mood disorders, and neurodegenerative diseases. Therefore, it is of great significance to study the pathogenic mechanism of sleep disorders and develop effective intervention methods to improve public health and reduce the burden of related diseases.

Research status at home and abroad:
In the central nervous system, the suprachiasmatic nucleus (SCN) is the core center of the mammalian biological rhythm, known as the "brain biological clock". Located above the hypothalamic optic chiasm, the SCN is composed of about 20,000 neurons that are able to autonomously generate a rhythm for about 24 hours and coordinate whole-body rhythms through nerve and humoral signals. SCN regulates almost all circadian rhythms, including sleep-wakefulness, such as eating, endocrine, body temperature, etc. Its main output is projected to several key regions of the hypothalamus, including the preoptic region (including the sleep promotion center VLPO), the dorsomedial hypothalamic nucleus (DMH), the paraventricular nucleus (PVN), and the thalamus and limbic system. Through these projections, SCNs can affect hypothalamic wakefulness-promoting regions (eg, orexin/amphetamine-related peptide neurons in the lateral hypothalamus) and the ascending brainstem arousal system, thereby promoting wakefulness or sleep at different phases of circadian rhythms. For example, the classical "flip switch" model proposes that during the subjective daytime period, SCN activates the wakefulness-promoting pathway and inhibits VLPO through the dorsomedial hypothalamus; When nocturnal SCN activity is reduced, VLPO is activated to induce sleep. SCN rhythm disturbances have been found to be closely associated with a variety of mood disorders and sleep disturbances. Decreased neural activity and output connections of SCN in older adults are thought to be an important cause of sleep-wake cycle disruption. Therefore, maintaining normal rhythmic signals is essential for sleep homeostasis. The study of SCN and its connection to central sleep function is the key to unraveling the mechanisms of rhythm-related sleep disorders.

In recent years, breakthroughs have been made in the analysis of sleep neural circuits, thanks to the development of optogenetics, chemogenetics, in vivo calcium imaging and other technologies. Optogenetics uses specific wavelengths of light to precisely control the excitation or inhibition of specific types of neurons, enabling researchers to modulate cellular activity in sleep circuits on the millisecond scale, thereby verifying the causal effect of a certain neuronal population on the sleep/wake transition. In the past decade, a large number of studies have dissected the neural circuits of sleep-wake regulation through optogenetic means, such as activating glutamatergic neurons in the lateral hypothalamus to induce wakefulness, and activating GABAergic neurons in VLPO to trigger sleep. Chemogenetics (e.g., DREADDs) can regulate neuronal activity for a long time through drug administration, which is convenient for studying circuit function changes in sleep deprivation, chronic insomnia, etc. Calcium imaging techniques such as in vivo two-photon imaging and fiber optic recording allow us to observe the dynamic activity patterns of sleep-related neuronal populations in the brains of free-moving animals in real time, and the correspondence with EEG and behavioral states, thereby revealing the cooperative activity characteristics of neural networks during sleep.

On the other hand, multiomics techniques are being widely used in sleep studies to find mechanisms and markers at the molecular level. Transcriptome sequencing, proteomic and metabolomic analyses showed that sleep deprivation led to extensive gene expression changes in the brain, and the differential regulation of different brain regions suggested that the molecular mechanism of sleep/wakefulness was region-specific. For example, some studies have used spatial transcriptome to find that transient sleep deprivation can cause significant changes in the expression of thousands of genes in multiple brain regions such as hippocampus, cortex, and hypothalamus, and the differential genes and regulatory directions in each region are different. In addition, proteomic and phosphoromic studies have revealed that sleep loss can trigger different protein pathway changes in astrocytes and neurons, involving processes such as synaptic plasticity, energy metabolism, and stress response. These multimodal omics data provide rich clues to our search for key molecules in sleep disorders. For example, abnormal day-and-night fluctuations in certain inflammatory factors and stress hormones in the hypothalamus and cerebrospinal fluid may be diagnostic markers for sleep rhythm disorders. For another example, the discovery of the orexin pathway has elucidated the mechanism of narcolepsy and directly led to the emergence of new drugs targeting this pathway, which is a model for the successful transformation of molecular mechanism research in the field of sleep.

It is worth noting that the recent research on the glymphatic system has provided a new perspective for understanding the restorative function of sleep. The glucolymphatic system is a waste product removal pathway in the brain that relies on glial cells to remove harmful metabolites through the convective exchange of cerebrospinal fluid between brain parenchyma during sleep. Studies have shown that more than 85% of metabolic waste removal occurs during sleep. During deep sleep, neuronal activity decreases and the extracellular space expands, which facilitates the removal of metabolites such as β-amyloid from cerebrospinal fluid lavage. If there is a lack of sleep or rhythm disorders, glucolymphatic function is impaired, which can lead to toxic protein deposition and accelerate the development of neurodegenerative diseases such as Alzheimer's disease. Therefore, the mechanism of action of sleep in metabolite clearance and brain health maintenance has become a hot topic. The identification of key molecules (such as AQP4 aquaporin, norepinephrine, etc.) that regulate sleep-clearance function through omics methods is of great significance for the development of new strategies for the prevention and treatment of neurodegenerative diseases.

The basis of the team's preliminary work:
Professor Yucong Duan, the applicant of this project, has deep research accumulation in the field of artificial intelligence and cognitive computing, and proposed the "DIKWP Artificial Consciousness Theory" to construct Data-Information-Knowledge-Wisdom-Purpose (DIKWP**). cognitive architecture. This theory expands the limitations of traditional artificial intelligence focusing on DIK (data, information, knowledge), and introduces higher-level Wisdom (W) and Purpose (P) processing units to form an artificial consciousness model. Professor Duan further proposed the network DIKWP ×DIKWP interaction model**, which is used to describe the interaction mechanism of multi-agent and multi-level intelligent systems, that is, the information exchange between different DIKWP units (agents) at all levels from data to the destination to form a dynamic network to achieve more complex cognitive and decision-making functions. This model can be compared to the network interaction between different functional regions of the brain, and has unique advantages in simulating sleep-wake brain circuit communication. Based on the DIKWP model, the team developed a prototype of the Artificial Consciousness Operating System (AC Operating System), which integrates the DIKWP system into the computing framework to realize the orchestration and management of cognitive processes. In conjunction with this OS, the team also explored semantic programming methods to directly manipulate AI systems in the form of natural language and knowledge graphs to perform complex analysis and decision-making tasks in accordance with human semantic purposes. This method is expected to be used for the integrated analysis of multi-source heterogeneous biomedical data, accelerating the transformation from data to knowledge/wisdom.

Although Professor Yucong Duan's original work is mainly focused on the field of artificial intelligence, his ideas have potential cross-border innovation value for the study of sleep rhythm regulation. In recent years, the applicant team has made positive progress in applying the DIKWP theory to biomedical informatics, such as using semantic network models to assist in the analysis of multi-omics data. In addition, the team has also accumulated experience in brain network modeling and biological signal processing, which provides technical support for the implementation of this project. Based on the previous work, we have the unique advantage of combining artificial consciousness models with neuroscience experiments, and can introduce advanced algorithms of artificial intelligence into the study of sleep disorder mechanisms in this project, and open up new research perspectives and technical means.

In summary, the research foundation of this project is sufficient: there is a solid theoretical and technical accumulation in the field of sleep and rhythm, and our team's original work in artificial intelligence and multimodal data fusion will inject new ideas into this project. Based on the above foundation, we are confident that we will make breakthroughs in the mechanism and intervention research of rhythm-disrupted sleep disorders, and provide innovative solutions to solve major scientific problems and clinical needs.

Research Objectives:

Focusing on the theme of "Pathogenesis and Intervention and Treatment of Sleep Disorders", the project aims to develop the following overall objectives:

• Elucidate the pathogenic mechanism of neural circuits in rhythmic disordered sleep disorders: construct animal models of biological rhythm disorders and sleep disorders, analyze the functional connectivity changes between the rhythmic center with the suprachiasmatic nucleus (SCN) as the core and sleep/wakefulness-related brain regions (such as VLPO, lateral hypothalamus, etc.), and reveal the circuit-level mechanism of rhythm disorders-induced sleep disorders.

• Revealing the key molecular mechanisms of sleep regulation and metabolic waste removal: Transcriptome, proteome, metabolome and other multi-omics technologies were used to systematically identify the key molecules of sleep-wake regulation and brain metabolite clearance during sleep. Elucidate how dysrhythmia affects these molecular pathways, leading to metabolic waste accumulation and neurological impairment, and screen for biomarkers that can be used in the diagnosis of sleep disorders.

• Innovative intervention and treatment strategies for non-invasive brain regulation: To study the improvement effect and mechanism of non-invasive brain regulation methods such as transcranial electrical/magnetic stimulation and light stimulation on sleep and rhythm disorders, and to explore the adaptive regulation mode based on brain-computer interface and artificial intelligence feedback. On this basis, new sleep regulation devices and therapies are developed, their safety and efficacy are evaluated, and intervention products or clinical programs with independent intellectual property rights are strodied.

• Integrating the theory of artificial consciousness to promote the innovation of sleep research: The applicant's original DIKWP artificial consciousness model is deeply integrated into the above research process, and a whole-process intelligent research platform from data collection to knowledge discovery, from neuromodulation to intelligent feedback is established. Through the artificial consciousness operating system and semantic programming, the semantic integration analysis of multi-modal data, the simulation modeling of sleep-wake brain circuits, and the intelligent optimization of control strategies are realized, so as to improve the research efficiency and innovation level, and form a leading interdisciplinary research paradigm.

Through the realization of the above goals, this project is expected to produce multi-level results such as brain functional connectivity map, key regulatory molecules, diagnostic markers, regulatory targets, and new treatment methods, which will provide a theoretical basis and technical reserve for the diagnosis, prevention and treatment of sleep disorders.

Scientific issues

Focusing on the research objectives, this project intends to focus on the following scientific questions:

1. Connectivity between the rhythm center and the sleep center and its disruptive effects: Under normal circumstances, how does SCN affect sleep/wake regulatory centers (e.g., VLPO, thalamus, brainstem reticular ascending system, etc.) through neural circuits to maintain sleep rhythm? When biorhythms are disrupted (e.g., circadian cycle disorders or clock gene dysfunction), how do the activity patterns and functional connections of these key circuits change, leading to abnormalities in sleep structure and behavior? What role do specific nerve nuclei and pathways play in rhythm-disrupted sleep disorders?

2. Sleep-driven removal of metabolic waste and its molecular regulation: How is the efficiency of the brain's removal of metabolic waste through the glue-lymphatic system regulated by rhythm signals during sleep? Does the dysrhythmia impair this clearance process? What are the molecular and cellular components that play a key role in sleep-clearance function (e.g., AQP4 water channels, norepinephrine, astrocytes, etc.), and how do their functional changes explain the cognitive decline and increased risk of neurodegeneration that are common in patients with sleep disorders? Is it possible to capture biomarkers in cerebrospinal fluid or blood that reflect the disorder of this process?

3. Mechanism of non-invasive brain stimulation on sleep rhythm: what are the effects and neurobiological mechanisms of non-invasive brain regulation technologies such as transcranial magnetic stimulation (rTMS), transcranial electrical stimulation (tDCS/TES), transcranial ultrasound, and ambient light regulation on improving sleep and correcting rhythm disorders? How do different stimulation patterns (frequency, intensity, timing) affect neural activity and network dynamics in sleep-related brain regions, thereby altering sleep states (e.g., prolonging the duration of deep sleep, improving sleep efficiency, etc.)? Are there individual differences and room for optimization in the role of these interventions, and how to achieve personalized and adaptive interventions?

4. Application of Artificial Intelligence and Artificial Consciousness in Sleep Research: How to Use the DIKWP Artificial Consciousness Model to Simulate and Predict the Complex Dynamics of the Sleep-Wake Process? Is it possible to integrate multi-source data (EEG, neural activity, omics, physiological signals, etc.) through the artificial consciousness operating system to realize intelligent monitoring and abnormal detection of sleep status? How does semantic programming play a role in mining the underlying mechanisms of sleep disorders and screening therapeutic targets? Finally, can an AI-assisted closed-loop feedback system be built to automatically adjust stimulation parameters according to real-time monitoring of sleep physiological indicators to achieve optimal regulation of sleep?

The above scientific questions are interrelated and cover the entire chain from the basic mechanism to the application transformation. Addressing these issues will enrich our systematic understanding of the occurrence and progression of sleep disorders and provide a scientific basis for the development of new therapies.

Research content and methodology

In order to answer the above scientific questions, the project has set up four major research contents, each part of which includes a number of specific research tasks and technical routes. Each part is connected to each other, layer by layer, and integrated into DIKWP artificial consciousness theory and related methods, so as to achieve the innovative integration of theory and technology.

1. Construction of model of rhythm disorder and sleep disorder and research on brain function connection

Main contents: To construct animal models of various rhythmic sleep disorders, to characterize their pathological characteristics from the macroscopic behavior to the microscopic neural circuit level, and to analyze the change map of brain functional connectivity with SCN as the core. This paper focuses on the changes in the strength and pattern of functional connectivity between SCN and sleep/wake regulation (VLPO, lateral hypothalamus LH, thalamus and brainstem reticular ascending system) under the condition of rhythm disorder.

Methods:

• Model establishment: A variety of methods were used to induce rhythm disorders and sleep disorders in mice. For example, disrupting the peripheral clock with ambient light interventions (inverting the circadian cycle, prolonging light, etc.); genetic knockout/mutation of core clock genes (e.g., Bmal1 or Clock); or apply jet lag simulation (changing the eating/activity cycle). Animal activity-resting rhythms and changes in sleep structure (EEG/EMG) were recorded to screen out models that stably reproduce the characteristics of human sleep disorders. Phenotypes of the model included decreased sleep duration, increased fragmentation, decreased circadian amplitude or phase dislocation.

• Brain network measurements: Multi-channel in vivo electrophysiological recordings combined with local field potentials of brain slices were used to measure the synchronization and information transfer between SCNs and downstream sleep-related nuclei in model animals and normal controls. Neural tracing techniques (anterograde and retrograde tracers) were used to map the anatomical connections of SCN to each node of the sleep circuit, and combined with c-Fos functional markers and fMRI functional imaging, to determine the degree of influence of rhythm disorders on the functional connections of these pathways. It is expected that a map of brain functional connectivity will reveal changes in the activity of key pathways (e.g., whether neural activity in the SCN-preoptic pathway is attenuated in the model, and whether VLPO is impaired in inhibitory function of the brainstem arousal center).

• Optogenetic and Fiber Optic Imaging Analysis: Select nodes with significantly altered functions in the above circuits, and use optogenetics to precisely regulate their activities to verify their causal role in sleep disorders. For example, in the arrhythmic disorder model, selective light stimulates the SCN output pathway to simulate normal rhythmic signals to see if it can correct the sleep disorder phenotype; or inhibiting overactive arousal to promote neuronal populations to see if it can improve sleep quality. At the same time, optical fiber recording/in vivo calcium imaging was used to monitor the dynamic activity of these neuronal populations at different stages of the sleep/wake cycle, so as to gain a deeper understanding of how rhythm signals regulate sleep occurrence. The data recorded by the fiber will be collected and analyzed in real time through the artificial consciousness OS: the "data-information" layer of the DIKWP model is used to filter and extract neural activity features, the "knowledge-wisdom" layer is used for pattern recognition and correlation analysis with sleep state, and the "purpose" layer is used to guide the next intervention (such as deciding when to give light stimulation to improve sleep).

• DIKWP network simulation: Based on the above experimental evidence, a mesh DIKWP × DIKWP interaction model was constructed to simulate the brain rhythm-sleep network. The SCN and the main target brain regions are regarded as agent nodes with DIKWP hierarchical processing architecture: the SCN nodes periodically generate rhythm signals (data layer); Downstream nodes receive and transform information, accumulate it into knowledge, and generate wisdom decisions on sleep/wakefulness (e.g., VLPO inhibits wakefulness center to induce sleep); Each node also has a goal to achieve a specific physiological purpose (homeostasis). The artificial consciousness operating system is used to orchestrate the interaction of these nodes to simulate the network dynamics under normal and disordered conditions. By comparing with the experimental observations, the model parameters were continuously adjusted to reconstruct the overall loop of sleep-rhythm regulation in the computer. This simulation can help predict the role of certain untested connections in the disorder, and provide hypothetical guidance for the next stage of experimentation.

Expected results: To obtain a visual map of the functional connectivity of sleep circuits under the condition of rhythm disorder, and identify 2~3 neural pathways (such as SCN-VLPO or SCN-LH pathway) that are critical for sleep disorders. The function of these pathways is verified by causal interventions. The loop simulation based on the DIKWP model will deepen the understanding of the systematic characteristics of the sleep loop and provide computational support for subsequent research.

2. Multiomics analysis of the molecular mechanism of sleep regulation and metabolic clearance

Main contents: In rhythm disorders and control animals, multi-level omics analysis of brain regions and body fluids related to sleep and metabolite clearance was performed to search for key molecules, pathways and biomarkers. Combining semantic programming and bioinformatics tools to transform multi-omics data into a systematic understanding of mechanisms.

Methods:

• Targeted tissue and temporal sampling: According to the results of study content 1, several key brain regions of sleep regulation (such as SCN, VLPO, lateral hypothalamus, hippocampus, etc.) and cerebrospinal fluid samples were selected. Tissue and body fluid samples were obtained from animals in different phases (subjective daytime, night, sleep and wake states) under normal circadian cycle and rhythm disturbances to capture dynamic change information. Particular attention is paid to differences between sleep and wakefulness, as well as abnormal patterns resulting from rhythm disturbances.

• Transcriptome and epimics: RNA sequencing (RNA-seq) was used to analyze the gene expression profile changes of the above brain regions under different conditions. It focuses on the differential expression of biorhythm genes (such as downstream genes regulated by Clock/Bmal1), sleep-related genes (such as encoding neurotransmitter receptors, ion channels, etc.), and metabolic and inflammatory pathway genes. Combined with epigeneomics (e.g., DNA methylation sequencing, histone modification ChIP-seq), this paper explores whether rhythm disorders affect the rhythmicity of gene expression through epiregulatory mechanisms. It is expected that dozens of candidate genes that are significantly dysregulated in sleep disorder models can be identified, providing clues for further research.

• Proteomics and metabolomics: High-resolution mass spectrometry is used to analyze changes in protein expression and modification in corresponding brain regions, as well as metabolite profiles in cerebrospinal fluid. Particular attention is paid to changes in the levels of proteins (e.g., AQP4, aquaporins, synaptic plasticity-related receptors, inflammatory factors, etc.) and metabolic wastes (e.g., lactate, β-amyloid fragments, etc.) associated with sleep-clearance function. Improve quantitation accuracy with labeled quantification (TMT or iTRAQ) and compare models to controls. Combined with metabolomic data, the disorders of energy metabolism, neurotransmitter metabolism and metabolic waste clearance pathways during sleep were revealed.

• Key molecule identification: Based on the results of multi-omics, 23 key regulatory molecules with repeated abnormalities in the sleep disorder model were screened. For example, it may be found that an inflammatory mediator and a hormone pathway molecule are consistently upregulated, suggesting that it may mediate a decrease in sleep quality. For these molecules, literature research and database analysis can be used to understand their function, and follow-up validation experiments (such as gene knockout or drug antagonism) can be designed to evaluate their causal effects. In addition, 12 potential diagnostic markers were screened from cerebrospinal fluid metabolites, which were required to show significant and detectable changes in the rhythm disorder model, which are expected to be detected clinically by non-invasive means for the diagnosis or risk prediction of sleep disorders in the future.

• Multimodal data integration and semantic analysis: In the face of massive omics data, we will introduce artificial intelligence to assist in integrated analysis. Using the semantic programming method developed by the applicant team, the knowledge graph related to sleep regulation was constructed, and the molecules and pathways found in the transcriptome, proteome, and metabolome were represented by semantic networks and associated with existing biological knowledge (such as Gene Ontology and KEGG pathways). Through the artificial consciousness OS, the data is processed in layers under the DIKWP framework: the data layer is standardized, the information layer is used for statistical analysis and pattern extraction (such as co-expression network and molecular module identification), the knowledge layer introduces literature evidence and biological pathway knowledge to explain these patterns, the Wisdom layer finds out the most likely regulatory core molecules through reasoning, and the purpose layer finally forms answers to research questions (such as determining the most critical pathological circuits and molecules). The whole process is controlled by semantic instructions, and the researchers ask the system in a way that is close to natural language (e.g., "find out the upstream regulators that cause sleep quality decline"), and the system synthesizes multiple sources of information through artificial consciousness models to provide answers or hypotheses. This will greatly improve the efficiency of the transformation of multi-omics data into biological knowledge.

• Validation experiments: Targeted validation of key molecules and markers identified. For example, immunohistochemistry was used to verify the expression changes of key proteins in model brain tissues; Brain stereotactic injection of specific receptor agonists/antagonists to target brain regions was used to observe the effect on sleep behavior; For markers, ELISA assays were designed to assess their level differences in more model individuals or pre-experimental clinical subjects to confirm their diagnostic value.

Expected results: 2~3 regulatory molecules and pathways that play a key role in arrhythmic sleep disorders were identified, and their mechanisms affecting sleep and metabolic clearance were clarified. These molecules will be potential targets for intervention. At the same time, 1~2 kinds of humoral markers that can be used for early diagnosis of diseases or efficacy monitoring were screened. Construct a molecular network model of sleep regulation with multi-layer information to deepen the systematic understanding of the molecular regulation mechanism of sleep. The efficient integration of multi-omics data with the help of the semantic analysis of the artificial consciousness framework is also an innovative methodological achievement in itself.

3. Research on the mechanism and mode of action of non-invasive brain regulation

Main contents: To study the effects of non-invasive brain stimulation and regulation methods on sleep and biorhythms, analyze their neural circuits and molecular mechanisms, optimize stimulation patterns, and explore personalized regulatory strategies. The purpose of this study is to provide a theoretical basis for the development of new intervention methods and lay a foundation for the next step of device development.

Methods:

• Screening of intervention methods: Several non-invasive brain modulation methods with current application prospects were selected, including transcranial magnetic stimulation (rTMS), transcranial electrical stimulation (tDCS/TES), transcranial low-intensity ultrasound stimulation, and audio-visual stimulation(e.g., sound pulses of specific frequencies, light pulses induce brain waves), etc. The effects of these measures on sleep markers were preliminarily tested in animal models of rhythm disorders. For example, model mice were subjected to rTMS with specific parameters during their active period, and the changes in EEG slow-wave activity, total sleep time, and number of awakenings during sleep were observed. or use slow-wave synchronized sound stimulation to enhance their deep sleep ratio during sleep. By comparing different methods and parameter combinations, 1~2 intervention methods with significant effects on improving the sleep of the model were screened.

• Mechanism of action research: In-depth analysis of the mechanism of action for the best methods screened. In animal experiments, a combination of electrophysiology and imaging techniques was used to investigate the immediate and long-term effects of stimuli on activity in relevant areas of the brain. For example, microelectrodes can be used to record changes in cortical and thalamic neurons when rTMS is applied, or functional imaging can be used to observe how thalamic reticular nucleus activity changes after low-intensity focused ultrasound stimulation, so as to infer the mechanism of its sleep-promoting neural circuitry. In the case of transcranial electrical stimulation, the modulation effect on the slow-wave oscillation of the cortex and its propagation effect to the downstream sleep center were recorded. The effects of these stimuli on neurotransmitter levels (e.g., promoting melatonin secretion or decreasing stress-related transmitters) will also be examined to assess their effects at the molecular level.

• Parameter optimization and mode exploration: Based on the understanding of the mechanism, the stimulation parameters are systematically optimized, such as adjusting the frequency and pulse sequence of rTMS to maximize the proportion of slow-wave sleep; Optimize tDCS electrode placement and current intensity to safely and effectively induce drowsiness. At the same time, rhythm-dependent stimulation patterns are explored: for example, in accordance with the circadian rhythm, sleep stimuli are given during the physiological sleep period, and intervention is avoided during the wake period to re-establish a normal rhythm. Artificial intelligence algorithms (reinforcement learning, etc.) are used to establish feedback models: sensors monitor the EEG and behavioral states of animals (or human volunteers) in real time, and AI decides when and in what mode to apply stimuli to achieve closed-loop adaptive regulation. The algorithm will be deployed on an artificial consciousness operating system, which continuously adjusts the stimulation program according to a set "purpose" (e.g., improving sleep quality) to simulate the autonomous optimization ability of artificial consciousness in the regulation process.

• Individualized Difference Analysis: Taking into account differences in the response of biological individuals to stimuli, this section will evaluate the feasibility of individualized regulation. For example, the intervention experiment was repeated on different animal individuals, and data on their underlying sleep lineage and response to stimuli were collected, and the classification of response patterns (e.g., "strongly responsive", "weakly responsive") was found through machine learning clustering. Combine multiomics or genetic analysis to explore the biological basis of individual differences (e.g., differences in stimulation effects due to polymorphisms in certain genes). This will help to implement individualized sleep intervention programs in the population in the future.

Expected results: To elucidate the mode and mechanism of action of non-invasive brain stimulation in improving sleep, for example, to demonstrate that rTMS can improve sleep quality by enhancing cortical-thalamic slow-wave network activity to prolong the duration of deep sleep. The combination of key parameters (such as optimal stimulation frequency, current intensity, etc.) and its neurobiological effects were identified to provide a basis for the subsequent formulation of clinical stimulation programs. A prototype of closed-loop brain-computer regulation principle was formed: a real-time monitoring and feedback stimulation system for sleep state based on artificial intelligence, and its effectiveness in animal models was verified. This achievement will lay the foundation for further transformation and application to the human body.

4. Development of new control devices and evaluation of treatment methods

Main content: On the basis of the above-mentioned mechanism research, a number of technologies are integrated to develop a new sleep disorder control device or system, and preliminary functional and safety evaluation is carried out. At the same time, new therapeutic approaches (including physical stimulation regimens and potential drug/biologics regimens) are designed to validate their efficacy and feasibility in animal models and preliminary clinical settings in combination with the previously obtained targets and molecules. Aim to produce an applicationable intervention product or preclinical outcome.

Methods:

• Development of intelligent sleep control device: Relying on the key technologies developed by this project, a prototype of an intelligent device integrating monitoring, analysis and intervention was designed. For example, the development of wearable headbands or cap devices that integrate multi-channel EEG acquisition and controllable stimulation modules (electrical or acoustic stimulation, etc.). The device is powered by a microprocessor or mobile terminal that runs an artificial consciousness OS that analyzes the user's sleep stages in real time and automatically triggers appropriate interventions (e.g., applying a mild stimulus to guide a return to deep sleep when a tendency to light sleep or arousal is detected) based on a built-in optimization algorithm (a closed-loop model from Study Content 3). The whole device emphasizes non-invasive, portable, and adaptive, and provides a human-machine interface for users to monitor sleep quality. The device will be tested under laboratory conditions, including algorithm accuracy (sleep staging discrimination is consistent with conventional polysomnography), reliability of stimulation output, and safety (e.g., current intensity is within a safe range with no significant side effects).

• Key technical research: In order to realize the above device functions, it is necessary to solve several technical problems, such as: long-term EEG acquisition with low noise and high comfort, miniaturized controllable stimulation circuit design, artificial consciousness OS transplantation and optimization in embedded platforms, and the application of semantic programming in human-computer interaction interface (enabling professional physicians to customize the intervention strategy of the device through high-level semantic instructions). In response to these, the project members will work together to tackle key problems, and use existing patents and experience. It is expected to form independent intellectual property rights in device hardware design, circuit safety, signal processing algorithms, etc.

• Design of new treatment methods: In addition to physical devices, this project also plans to propose 1~2 innovative treatment ideas based on the discovery of new targets/molecules. For example, if omics studies have found that inflammatory pathways play a prominent role in rhythm disorders, an evaluation of the effect of anti-inflammatory drugs in animal models may be tried; Or for comorbidities such as depression associated with rhythm disorders, explore new schemes of combined psychobehavioral therapy. In addition, noninvasive stimulation can be combined with medications (e.g., low-dose melatonin during the day to adjust the phase, and brain stimulation at night to prolong deep sleep) to form a comprehensive intervention strategy. Each candidate will be validated in animal models for efficacy, such as improving sleep duration, restoring circadian rhythms, etc., and evaluating the safety window. Excellent protocols will be screened for further translational research.

• IIT Studies and Preliminary Clinical Trials: Later in the program, if a device or therapy proves to be effective and safe in animals, we will initiate investigator-initiated clinical trial (IIT) preparation. After ethics approval, recruit small-scale subjects (e.g., 10-20 volunteers with sleep disorders) to try out new devices or protocols under strict monitoring. The observed indicators included subjective sleep quality score, objective polysomnography parameters, and daytime function improvement. Trials will assess the effects of the intervention using a randomised controlled or self-controlled before-and-after design. At the same time, improve the layout of intellectual property rights (such as timely application for invention patents and software copyrights) to clear the obstacles for subsequent large-scale clinical application and industrial transformation. Strive to complete 1~2 IIT studies with independent intellectual property rights, and push the laboratory results to the substantive stage of clinical transformation.

Expected results: 2~3 innovative and validated methods for the regulation of sleep disorders were developed. This may include a prototype device for intelligent closed-loop sleep regulation, as well as several optimized stimulation therapies or drug combinations. Through performance tests and preliminary trials, it has been proven to be effective and safe in improving sleep and regulating rhythms. Apply for a number of patents, publish guidelines or specification suggestions, and lay the foundation for future product registration and promotion. The results of this section will mark a critical leap from mechanistic research to applied translation.

Technical route

In order to ensure the smooth implementation of the above research contents, this project formulates a clear technical roadmap, which is promoted from basic to application, and each stage supports and verifies each other:

1. Model Establishment and Circuit Interpretation Phase (Year 1–2): First, an animal model of rhythmic sleep disorder is constructed to obtain behavioral and basic physiological characteristics. Subsequently, brain circuit studies were performed on the model, including neural tracing, electrophysiological recording, and optogenetic interventions, functional connectivity mapping and identification of key abnormal circuits. The artificial consciousness model was used to simulate the loop to form a preliminary hypothetical framework for the mechanism.

2. Multi-omics Integration and Mechanism Mining Phase (Year 2–3): Multi-omics data is collected from model and control animals, and a combination of experimental and artificial intelligence analysis is used to extract a list of molecules and pathway networks associated with sleep and metabolic clearance disorders. The multi-omics results can be used to explain the discovery of the loop level (e.g., molecular changes in a pathway support an abnormality in a circuit) and to provide new targets for intervention on the other. During this period, the data is integrated through semantic programming, and the scope of key problems is gradually narrowed.

3. Intervention Validation and Optimization Phase (Year 3–4): For the loops and molecular targets proposed in the first two stages, non-invasive stimulation and other means are used to verify the intervention. By continuously adjusting the stimulation method and parameters, and optimizing with the help of AI feedback, the sleep phenotype of the model was improved. The verification results will be fed back to the models and mechanisms in the previous stage to improve the understanding of pathological mechanisms. At the same time, the mechanism of action of the intervention means is studied in depth at this stage in order to optimize its effect.

4. Device Development and Initial Transformation Phase (Year 4–5): Design and develop intelligent control devices and new treatment options based on all previous results. After the small-scale trial is effective, the IIT research will be promoted. Improve technical details to form standard operating procedures and preclinical data. At the end of the project, we will strive to achieve a closed loop of transformation from mechanism research to application prototype.

Throughout the above phases, the AI-aware DIKWP framework will serve as a unified platform to support data analysis, model building, and feedback control. This technical approach ensures that we can apply the results to the design of intervention programs in a timely manner while exploring basic scientific problems, and finally achieve a comprehensive breakthrough from theory to practice.

Organization & Management

This project is led by Prof. Yucong Duan and is based on his unit to form an interdisciplinary research team. The team members include experts and key young researchers in the fields of neuroscience, molecular biology, artificial intelligence and computer science, forming a complementary and collaborative organizational structure. The specific organization and management plan is as follows:

• Division of labor and team structure: The project has set up one overall person in charge (PI), namely Professor Yucong Duan, to comprehensively coordinate the implementation of the project. There are several project module leaders: for example, 1 person in charge of neural circuit project, responsible for animal model and brain circuit research; 1 person in charge of molecular mechanism project, responsible for multi-omics experiments and analysis; 1 person in charge of technical device project, responsible for project development and transformation; 1 person in charge of the artificial intelligence project, responsible for the DIKWP model integration and data analysis platform. The person in charge has extensive experience in the corresponding field. The team also includes several postdoctoral fellows and graduate students, who are involved in the specific experiments and development of each module. In terms of organizational structure, a project coordination team is set up, composed of PI and the person in charge of each module, to communicate progress and coordinate resources on a regular basis.

• Planning & Milestone Management: Develop a detailed research plan and timeline that divides the five-year research cycle into milestones. Hold a project team meeting every six months to report progress and problems. Milestones include: completion of the model and preliminary loop map by the end of the first year; At the end of the second year, omics data collection and key molecule screening were completed; At the end of the third year, the verification of the main intervention means was completed; At the end of the fourth year, the prototype of the device was completed; At the end of the fifth year, the preliminary study and project summary of IIT will be completed. Each milestone is led by the corresponding person in charge and supervised and evaluated by the PI.

• Communication and coordination mechanism: The project team will maintain close communication through monthly meetings and symposiums. The experimental data and analysis results are shared in the form of online + offline combination. Establish a cloud-based project database and collaboration platform to save all raw data, analysis code, and stage reports, so that team members can access and update them in real time. Cross-training is encouraged in the project, such as allowing people with computer backgrounds to participate in biological experiment observation, and people with biological backgrounds to learn AI analysis skills to enhance the overall team's ability to collaborate.

• Risk control and contingency plan: Formulate alternative plans in advance for the challenges that may arise in each research process. If it is difficult to construct animal models, multiple model strategies should be used in parallel (environmental and genetic). If a key technology such as optogenetics or omics platform fails, a joint laboratory or commercial service support will be sought. The project has set up a risk monitoring specialist (served by an experienced associate researcher) to regularly assess progress risks and technical risks, report to the PI in a timely manner and initiate the plan. Ensure that the project progresses as planned, and important nodes do not drop the chain.

• Funding management and resource guarantee: In strict accordance with the key project fund management measures, the PI is responsible for the overall budget allocation, and the person in charge of each project specifically manages the part of the funds. The use of funds is linked to the scientific research plan and is reviewed regularly. The unit will provide support such as experimental sites, animal rooms, and instrument sharing platforms. The team will also actively strive for supporting resources inside and outside the university, such as hospital clinical collaboration support, corporate technology sponsorship, etc., to form a resource synergy.

• Achievement management and sharing: The data and results generated by the project are managed in strict accordance with the requirements of scientific research integrity. Apply for patents or publish papers in a timely manner for important discoveries, and clarify the authorship and ownership of intellectual property rights. External release follows the confidentiality and release approval system to protect the core technology that has not yet been disclosed. At the same time, the project encourages open science, and gradually discloses some data and tools (such as publishing semantic analysis codes and omics datasets) under the premise of ensuring intellectual property rights, so as to promote academic exchanges and downstream research.

Through the above organizational and management measures, the project team is ensured to collaborate efficiently, make full use of resources, and complete research tasks according to quality and quantity. Professor Yucong Duan's rich experience in project management and team cohesion will provide a strong guarantee for the smooth implementation of the project.

Expected outcomes and conversion paths

Focusing on the pathogenic mechanism and intervention and treatment of sleep disorders, this project is expected to produce the following multi-level results, and has developed a clear path for the transformation of results:

• Basic scientific achievements: A detailed map of brain functional connections was drawn to show the connection between the rhythm center SCN and the sleep-wake center and the changes under rhythm disorders. This atlas will deepen the understanding of the architecture of the sleep circuit and can be published in academic journals. At the same time, 2**–3 key regulatory molecules were identified** to elucidate their mechanisms of action in sleep regulation and metabolite clearance. Screening for 1**–2 diagnostic markers** that can be used to objectively assess rhythm-related sleep disorders. The above findings will be published in the form of a paper in a top journal of neuroscience or sleep medicine, and the corresponding invention patent will be applied for to protect the promising molecular marker combination or detection method.

• Technological Invention and Patents: This project will identify 2–4 potential regulatory targets (including neural circuit nodes and molecular targets). The development of novel interventions for these targets, such as specific brain stimulation parameter regimens and the application of new drug targets, is expected to lead to 2**–3 feasible therapeutic strategies. For the core device prototype and algorithm, we will apply for patents (such as "sleep regulation system and its method based on artificial consciousness") and software copyright, and realize the layout of independent intellectual property rights. These technological inventions can be further optimized into products in the future.**

• Cross-cutting achievements of artificial intelligence: The research platform that deeply integrates the DIKWP artificial consciousness model will be one of the important achievements of this project. Through the implementation of the project, we will develop a set of multimodal biological data semantic integration and intelligent decision-making system to verify its function in sleep studies. This system and its application in sleep regulation will also be published in a paper and patented. Its significance is not only to serve this project, but also to promote data analysis and intelligent intervention in other biomedical fields.

• Platform & Talent: To build a comprehensive platform for sleep disorder mechanism and intervention research, integrating animal models, omics analysis, brain stimulation technology, and AI algorithms. The platform can continue to be used for subsequent research on a larger sample and a wider range of topics. The project will also cultivate a group of interdisciplinary talents, including young researchers who have mastered neuroscience experiments and understand artificial intelligence analysis, so as to improve China's talent pool in this interdisciplinary field.

• Clinical translation pathway: For the screened diagnostic markers, we will cooperate with the hospital laboratory department to develop detection kits to carry out clinical sample validation, and strive to incorporate them into clinical evaluation methods. For the identified intervention targets, if there are ready-made marketed drugs, the indication external use trial can be carried out; If new drug development is required, pharmaceutical companies are sought to cooperate in the design of lead compounds. For the smart device developed in this project, after the completion of the preliminary verification of IIT, the productization will be improved, the user interface and safety certification will be improved, and the medical device approval process will be entered within 1-2 years after the end of the project. We plan to connect with medical technology companies to seek industrialization cooperation and bring the device to the market to serve the majority of patients with insomnia and rhythm disorders. This will be achieved through technology transfer or industry-university-research co-construction, and is expected to enter the clinical promotion stage in 5 years.

• Milestone products: During the implementation of the project, corresponding results will be produced every year: for example, the preliminary research paper on sleep circuit will be published in the 1st and 2nd years, the omics and key molecule papers will be published in the 3rd year, and the intervention therapy and device development paper will be published in the 4th and 5th years. At the same time, he participates in academic conferences at home and abroad every year to report progress and improve influence. In the end, more than 5 high-level papers, more than 5 invention patents, and a number of soft works and standards were formed. In particular, with the support of key projects, we strive to break through key technologies and form practical results that can be promoted, which is an important indicator to evaluate the success of the project.

In summary, the expected outcomes of this project are rich and pragmatic, covering the complete innovation chain from theoretical models to concrete products. We have designed a clear transformation route to ensure that these achievements can move from paper patents to clinical application and industrial development, realize the closed loop of scientific discovery, technological innovation and clinical application, and bring tangible benefits to patients with sleep disorders.

Social & Industrial Impact

The study of the pathogenic mechanism and intervention treatment of sleep disorders is of great social and industrial significance. The success of this project will have a positive impact in a number of ways:

• Improving the health level of the whole people: Sleep disorders, especially insomnia and rhythm disorders, have become prominent problems affecting the physical and mental health of the public, and long-term sleep deprivation can lead to decreased work efficiency, increased risk of accidents, and induce a series of diseases such as depression and hypertension. Through this project, we will deeply uncover the mechanism of sleep disorders and develop effective interventions, which are expected to significantly improve the sleep quality and quality of life of patients. This will reduce medical expenditures and socio-economic losses caused by sleep problems, which is in line with the goal of improving people's health in the "Healthy China 2030" plan, and will have a far-reaching impact on social health.

• Serving an aging society and cognitive health: With the aging of the population, sleep disorders and the resulting decline in cognitive function and the increased risk of dementia in the elderly are becoming increasingly prominent. The project focuses on the mechanism of brain metabolic waste removal during sleep, and may discover new ways to improve brain waste removal and prevent Alzheimer's disease. If the neurodegenerative process can be delayed, it will greatly reduce the burden of elderly care and medical insurance. The exploration of this project in this aspect has important social benefits.

• Promote innovation in related industries: The sleep health and brain science industry is a new growth point in the field of biomedicine and artificial intelligence in the future. The technical achievements of this project (such as intelligent sleep control devices and diagnostic reagents) will give rise to new products and services once they mature. For example, smart sleep headband devices can enter the wearable device market and digital health to meet people's needs for self-health management; New biomarker detection can be used for physical examination and chronic disease management, and improve the level of related in vitro diagnostic industry. It is expected to form a potential market of hundreds of millions of yuan, drive the development of upstream and downstream enterprises, and promote industrial upgrading.

• Leading the integration and innovation of artificial intelligence and brain science: This project is a bold cross-border attempt to apply artificial consciousness theory to brain science research. Its success will set an example for the application of AI in the field of life sciences and promote the development of AI technology to a higher level (brain-like intelligence, cognitive computing). This will enhance China's international influence in the interdisciplinary research of artificial intelligence. In terms of talent training, the interdisciplinary scientific research talents cultivated by this project will become a valuable resource for the academic and industrial circles, and further consolidate China's leading position in the fields of brain-computer interface and next-generation AI.

• Environmental and social cognitive impacts: The scientific value of the project should not be overlooked. We will disseminate scientific work and sleep health knowledge to the public through media reports and popular science lectures, raise the awareness of the importance of sleep among the whole people, and form a healthy lifestyle and social atmosphere. This will indirectly reduce social problems caused by drowsy driving, industrial accidents, etc., and promote social civilization and safety.

In short, the implementation of this key project will produce major scientific and technological achievements, directly serve the health and well-being of the people, and indirectly promote the development and social progress of related high-tech industries. Its impact will extend from medicine to the economic and socio-cultural spheres, with significant comprehensive benefits. We are confident that through this project, we can achieve a win-win situation of scientific innovation and social value, and contribute to the construction of China's scientific and technological power and a healthy China.