Description:FIELD OF DISCLOSURE
[0001] The present disclosure relates generally relates to the field of neurotechnology, computational neuroscience, and behavioral biometrics. More specifically, it pertains to a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models.
BACKGROUND OF THE DISCLOSURE
[0002] The pursuit of reliable methods to detect deception has been a longstanding challenge across psychology, neuroscience, law enforcement, national security, and interpersonal communication. Human beings have engaged in lying since the earliest stages of social development, and deception has become embedded in many aspects of human interaction. Traditional methods of detecting lie, such as behavioral observation and questioning strategies, have often proved insufficient because deception can be subtle, culturally variable, and context-dependent. Over time, this inadequacy has driven research towards integrating cognitive science, psychophysiology, and computational intelligence to develop objective and quantifiable lie detection mechanisms.
[0003] Early approaches to lie detection were heavily reliant on physical cues observable to the naked eye. Observers would interpret microexpressions, eye movements, vocal hesitations, or body posture changes as potential indicators of deceptive behavior. However, such methods have consistently suffered from subjectivity and inconsistency. Cultural differences further complicated the issue, as gestures or facial expressions considered deceptive in one culture could be perfectly normal in another. Researchers began to realize that deception is not simply a behavioral performance but is deeply rooted in the interplay of cognition, emotion, and neurophysiological activity. This realization prompted the transition from behavioral methods to psychophysiological techniques that could offer a more scientific foundation.
[0004] The polygraph records physiological signals such as heart rate, blood pressure, respiration, and skin conductance under the assumption that lying induces stress, which in turn produces measurable changes in these bodily functions. While polygraphs have been widely used in forensic investigations, employment screenings, and even intelligence interrogations, they have long been criticized for their questionable accuracy. Numerous studies have shown that polygraph results can be influenced by anxiety, fear, or medical conditions unrelated to deception. Furthermore, trained individuals may learn countermeasures to manipulate their physiological responses, further undermining reliability. Despite these limitations, polygraphs established the groundwork for exploring the link between mental states and bodily signals in lie detection research.
[0005] Studies using fMRI revealed that deception activates specific brain regions associated with executive control, working memory, and inhibition, particularly in the prefrontal cortex and anterior cingulate cortex. These findings lent strong support to the cognitive load theory of deception, which posits that lying is more mentally demanding than truth-telling because it requires fabricating information while simultaneously suppressing the truth. However, while fMRI studies contributed to a deeper understanding of the neural basis of lying, the technique suffered from practical challenges. fMRI is expensive, requires a controlled laboratory environment, and is impractical for field or forensic settings where portability and real-time analysis are essential.
[0006] Electroencephalography (EEG) subsequently emerged as a more practical and cost-effective tool for investigating deception. EEG measures electrical activity of the brain with high temporal resolution, allowing researchers to track rapid neural changes associated with cognitive and emotional processes during deception. Unlike fMRI, EEG is relatively inexpensive, portable, and capable of delivering real-time insights. Early EEG-based lie detection methods often focused on event-related potentials (ERPs), particularly the P300 wave. The P300 is a positive deflection in brain activity occurring around 300 milliseconds after stimulus presentation and is thought to reflect recognition and attention. In so-called “guilty knowledge tests,” subjects are presented with stimuli related to concealed information, and the presence of a P300 response suggests recognition even if the subject denies knowledge.
[0007] While ERP-based EEG methods improved upon polygraph reliability, they too faced challenges. The P300 response can be influenced by attention, fatigue, or unrelated memory processes, which can lead to false positives or false negatives. Furthermore, deception is not merely a matter of recognition but involves complex interactions between emotion regulation, cognitive control, and moral decision-making. Researchers began to acknowledge that single-dimensional approaches focusing on either cognition or physiology were insufficient. A deeper, integrative perspective that accounts for both emotional and cognitive dynamics became necessary.
[0008] Parallel to advances in neuroscience, computational methods for analyzing complex physiological and behavioral signals also advanced. Machine learning techniques began to be applied to EEG data for deception detection, allowing algorithms to identify subtle patterns invisible to human analysis. Traditional machine learning models such as support vector machines, decision trees, and k-nearest neighbors demonstrated potential for classifying truthful versus deceptive responses. However, these models often required handcrafted features and could not easily capture the temporal dynamics of deception-related brain activity.
[0009] The rise of deep learning in the past decade provided new opportunities for EEG-based lie detection. Deep neural networks, including convolutional neural networks (CNNs) and recurrent neural networks (RNNs), have been increasingly applied to EEG analysis, offering the ability to automatically extract hierarchical features from raw signals. CNNs are particularly effective at capturing spatial features across EEG channels, while RNNs, especially long short-term memory (LSTM) networks, are adept at modeling temporal dependencies. These capabilities allow for a richer and more nuanced understanding of the complex spatiotemporal dynamics underlying deception.
[0010] A crucial realization in deception research is that lying is not purely cognitive but also deeply emotional. Lying often induces stress, guilt, or anxiety, all of which manifest in emotional processing regions of the brain as well as autonomic nervous system activity. Conversely, skilled liars may suppress these emotional responses or even experience excitement when deceiving. This dual involvement of cognition and emotion suggests that lie detection systems must account for the synchrony between cognitive control mechanisms and emotional regulation processes. Neurobehavioral studies have shown that cognitive load and emotional arousal often interact during deception, producing distinctive patterns across EEG and other biosignals. These findings have driven research into multi-task and multimodal systems that integrate emotion-cognition synchrony as a basis for more robust lie detection.
[0011] The evolution of explainable artificial intelligence (XAI) has also had an impact on lie detection research. Traditional deep learning systems, while powerful, often function as black boxes, offering little insight into how decisions are made. In high-stakes domains such as legal proceedings, national security, or clinical assessment, the interpretability of lie detection outcomes is critical. XAI techniques such as attention mechanisms, feature attribution methods, and saliency mapping have been increasingly applied to EEG and physiological data to reveal which neural patterns or time segments contribute most to deception classification. By combining accuracy with interpretability, XAI enhances trust in automated lie detection systems and enables human experts to validate findings.
[0012] Beyond neuroscience and AI, broader socio-legal and ethical considerations shape the field of lie detection research. The admissibility of polygraph evidence in court has long been contested due to questions of scientific validity. Similarly, the use of neuroimaging or EEG-based lie detection raises questions about privacy, consent, and potential misuse. Some critics argue that compelling individuals to undergo brain-based lie detection infringes upon cognitive liberty the right to control access to one’s thoughts. Others worry about the potential for false accusations if systems produce erroneous results. As such, the development of reliable, ethically sound, and legally defensible lie detection systems remains a pressing challenge.
[0013] At the same time, the practical applications of lie detection continue to expand. In law enforcement, investigators seek tools to identify suspects who may be withholding critical information. In counterterrorism and national security, lie detection technologies are viewed as potential aids in screening individuals at borders or during interrogations. In corporate contexts, organizations may use deception detection for fraud prevention or insider threat mitigation. In clinical psychology, understanding deception can assist in diagnosing certain disorders associated with compulsive lying or impaired social cognition. Each of these domains presents unique requirements, including accuracy, portability, scalability, and transparency.
[0014] Recent years have witnessed increasing interest in multi-task learning approaches, in which a single model is trained to perform multiple related tasks simultaneously. Applied to lie detection, multi-task models may simultaneously predict truthfulness, cognitive load, and emotional arousal, thereby capturing the multifaceted nature of deception. Such approaches leverage shared representations between tasks, improving generalization and robustness. Coupled with attention mechanisms, which allow models to focus on the most relevant temporal or spatial features in EEG data, multi-task architectures represent a promising frontier in neurobehavioral lie detection research.
[0015] Another growing area of interest is multimodal fusion, where EEG signals are combined with other physiological or behavioral data such as facial expressions, eye tracking, heart rate variability, or speech. These multimodal systems aim to capture the full spectrum of deception-related signals across cognitive, emotional, and behavioral domains. While promising, multimodal approaches also raise challenges related to synchronization, feature integration, and computational complexity. Nonetheless, the pursuit of holistic models reflects the recognition that deception is a multi-dimensional phenomenon that cannot be reduced to a single neural or behavioral marker.
[0016] The trajectory of lie detection research reflects a gradual but profound shift from surface-level observations toward deeper, mechanistic insights into the neurocognitive underpinnings of deception. What began as reliance on visible behavioral cues evolved into physiological monitoring, then into direct brain-based measurement, and now into sophisticated AI-driven models capable of decoding complex neural dynamics. Alongside these technological advances, the theoretical understanding of deception has matured, moving from simplistic notions of stress-induced physiological change to nuanced models of cognitive-emotional synchrony, executive control, and neural network interactions.
[0017] Yet despite these advances, the problem of lie detection remains unresolved. Each generation of methods has brought improvements but also limitations. Polygraphs lack specificity, fMRI lacks practicality, ERP methods lack robustness, and even modern machine learning systems face challenges of generalizability, interpretability, and ethical acceptance. This persistent gap underscores the need for continued innovation that integrates neuroscience, artificial intelligence, psychology, and ethics in a coherent framework. The pursuit of accurate, interpretable, and ethically defensible lie detection systems remains a grand challenge at the intersection of human behavior and technology.
[0018] Thus, in light of the above-stated discussion, there exists a need for a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models.
SUMMARY OF THE DISCLOSURE
[0019] The following is a summary description of illustrative embodiments of the invention. It is provided as a preface to assist those skilled in the art to more rapidly assimilate the detailed design discussion which ensues and is not intended in any way to limit the scope of the claims which are appended hereto in order to particularly point out the invention.
[0020] According to illustrative embodiments, the present disclosure focuses on a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models which overcomes the above-mentioned disadvantages or provide the users with a useful or commercial choice.
[0021] An objective of the present disclosure is to integrate a cognitive calibration phase for establishing personalized neural baselines, thereby improving detection accuracy and addressing the variability across individuals that earlier models overlooked.
[0022] Another objective of the present disclosure is to design a real-time, low-latency machine learning pipeline capable of providing instant deception analysis, overcoming the limitations of traditional EEG-based lie detection systems that often operate offline.
[0023] Another objective of the present disclosure is to develop a robust EEG-based lie detection system that analyzes direct brain activity rather than peripheral physiological signals, thereby making the system more resistant to manipulation compared to conventional polygraph techniques.
[0024] Another objective of the present disclosure is to implement a countermeasure detection module that identifies and flags intentional efforts to manipulate results, such as controlled breathing, forced eye fixation, or other behavioral artifacts.
[0025] Another objective of the present disclosure is to harness emotion-cognition synchrony as a neurobehavioral marker for deception, capturing the interplay between affective and cognitive processes to enhance reliability beyond single-domain analysis.
[0026] Another objective of the present disclosure is to employ deep attention-based neural architectures that dynamically focus on critical EEG features and temporal dependencies, thus improving sensitivity to subtle deception-related patterns.
[0027] Another objective of the present disclosure is to incorporate adaptive learning mechanisms that enable the model to evolve and recalibrate during ongoing sessions, providing robustness against shifts in cognitive or emotional states.
[0028] Another objective of the present disclosure is to ensure transparency and interpretability of model predictions by applying explainable AI techniques, thereby supporting admissibility and trust in legal, forensic, and clinical applications.
[0029] Another objective of the present disclosure is to create a multi-task learning framework that simultaneously handles deception detection, emotional state recognition, and cognitive load estimation, leveraging shared EEG features to improve generalization.
[0030] Yet another objective of the present disclosure is to validate the system in decision-making contexts such as forensic investigations, legal assessments, and clinical diagnostics, ensuring that it provides practical, ethical, and scientifically grounded support.
[0031] In light of the above, a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models comprises a multi-channel EEG data acquisition module configured to record brain activity signals associated with a subject’s responses in real-time. The system also includes a dual-branch deep learning processing unit comprising a cognitive analysis branch and an emotional analysis branch. The system also includes a spatiotemporal attention module configured to dynamically identify and emphasize EEG signal features from relevant brain regions and critical time intervals. The system also includes a synchrony fusion module configured to integrate outputs from the cognitive analysis branch and the emotional analysis branch to capture interactions between cognitive and emotional responses during deceptive behavior. The system also includes a classification module configured to generate deception likelihood scores and corresponding confidence levels based on the fused outputs. The system also includes an explainable decision interface configured to provide interpretable outputs highlighting contributing neural features and time windows relevant to deception classification. The system also includes a portable deployment unit comprising hardware and software elements that enable real-time operation of the system for forensic, clinical, or security applications.
[0032] In one embodiment, the dual-branch deep learning processing unit is trained using labeled datasets comprising truthful and deceptive responses, thereby enabling supervised learning of deception-related neural patterns.
[0033] In one embodiment, the cognitive analysis branch of the dual-branch deep learning processing unit is specifically configured to detect workload, hesitation, or mental conflict through analysis of frequency band activity within frontal and parietal EEG regions.
[0034] In one embodiment, the emotional analysis branch of the dual-branch deep learning processing unit is configured to identify stress, guilt, or arousal-related neural signatures by monitoring event-related potentials and power spectral densities.
[0035] In one embodiment, the spatiotemporal attention module applies weighted feature selection to dynamically prioritize temporal segments and spatial channels of EEG data most relevant for deception detection.
[0036] In one embodiment, the classification module generates deception likelihood scores using probabilistic inference models, recurrent neural networks, or support vector machines trained on fused synchrony features.
[0037] In one embodiment, disturbances in synchrony patterns detected by the synchrony fusion module are utilized to identify countermeasures including emotional suppression, rehearsed control, or cognitive distraction techniques.
[0038] In one embodiment, the explainable decision interface provides natural language summaries to enable forensic, clinical, or legal experts to interpret the system’s outputs without requiring specialized knowledge of neural signal processing.
[0039] In one embodiment, the portable deployment unit comprises wearable EEG headsets integrated with wireless communication modules, enabling real-time remote deception assessment.
[0040] In one embodiment, the portable deployment unit further comprises cloud-based storage and computation facilities to support large-scale deployment across forensic and security networks.
[0041] These and other advantages will be apparent from the present application of the embodiments described herein.
[0042] The preceding is a simplified summary to provide an understanding of some embodiments of the present invention. This summary is neither an extensive nor exhaustive overview of the present invention and its various embodiments. The summary presents selected concepts of the embodiments of the present invention in a simplified form as an introduction to the more detailed description presented below. As will be appreciated, other embodiments of the present invention are possible utilizing, alone or in combination, one or more of the features set forth above or described in detail below.
[0043] These elements, together with the other aspects of the present disclosure and various features are pointed out with particularity in the claims annexed hereto and form a part of the present disclosure. For a better understanding of the present disclosure, its operating advantages, and the specified object attained by its uses, reference should be made to the accompanying drawings and descriptive matter in which there are illustrated exemplary embodiments of the present disclosure.
BRIEF DESCRIPTION OF THE DRAWINGS
[0044] To describe the technical solutions in the embodiments of the present disclosure or in the prior art more clearly, the following briefly describes the accompanying drawings required for describing the embodiments or the prior art. Apparently, the accompanying drawings in the following description merely show some embodiments of the present disclosure, and a person of ordinary skill in the art can derive other implementations from these accompanying drawings without creative efforts. All of the embodiments or the implementations shall fall within the protection scope of the present disclosure.
[0045] The advantages and features of the present disclosure will become better understood with reference to the following detailed description taken in conjunction with the accompanying drawing, in which:
[0046] FIG. 1 illustrates a flowchart outlining sequential step involved in a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models, in accordance with an exemplary embodiment of the present disclosure;
[0047] FIG. 2 illustrates a flowchart showing working of a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models, in accordance with an exemplary embodiment of the present disclosure.
[0048] Like reference, numerals refer to like parts throughout the description of several views of the drawing;
[0049] The multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models, which like reference letters indicate corresponding parts in the various figures. It should be noted that the accompanying figure is intended to present illustrations of exemplary embodiments of the present disclosure. This figure is not intended to limit the scope of the present disclosure. It should also be noted that the accompanying figure is not necessarily drawn to scale.
DETAILED DESCRIPTION OF THE DISCLOSURE
[0050] The following is a detailed description of embodiments of the disclosure depicted in the accompanying drawings. The embodiments are in such detail as to communicate the disclosure. However, the amount of detail offered is not intended to limit the anticipated variations of embodiments; on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the present disclosure.
[0051] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. It may be apparent to one skilled in the art that embodiments of the present disclosure may be practiced without some of these specific details.
[0052] Various terms as used herein are shown below. To the extent a term is used, it should be given the broadest definition persons in the pertinent art have given that term as reflected in printed publications and issued patents at the time of filing.
[0053] The terms “a” and “an” herein do not denote a limitation of quantity, but rather denote the presence of at least one of the referenced items.
[0054] The terms “having”, “comprising”, “including”, and variations thereof signify the presence of a component.
[0055] Referring now to FIG. 1 to FIG. 2 to describe various exemplary embodiments of the present disclosure. FIG. 1 illustrates a flowchart outlining sequential step involved in a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models, in accordance with an exemplary embodiment of the present disclosure.
[0056] A multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models 100 comprises a multi-channel EEG data acquisition module 102 configured to record brain activity signals associated with a subject’s responses in real-time.
[0057] The system also includes a dual-branch deep learning processing unit 104 comprising a cognitive analysis branch and an emotional analysis branch. The dual-branch deep learning processing unit 104 is trained using labeled datasets comprising truthful and deceptive responses, thereby enabling supervised learning of deception-related neural patterns. The cognitive analysis branch of the dual-branch deep learning processing unit 104 is specifically configured to detect workload, hesitation, or mental conflict through analysis of frequency band activity within frontal and parietal EEG regions. The emotional analysis branch of the dual-branch deep learning processing unit 104 is configured to identify stress, guilt, or arousal-related neural signatures by monitoring event-related potentials and power spectral densities.
[0058] The system also includes a spatiotemporal attention module 106 configured to dynamically identify and emphasize EEG signal features from relevant brain regions and critical time intervals. The spatiotemporal attention module 106 applies weighted feature selection to dynamically prioritize temporal segments and spatial channels of EEG data most relevant for deception detection.
[0059] The system also includes a synchrony fusion module 108 configured to integrate outputs from the cognitive analysis branch and the emotional analysis branch to capture interactions between cognitive and emotional responses during deceptive behavior. Disturbances in synchrony patterns detected by the synchrony fusion module 108 are utilized to identify countermeasures including emotional suppression, rehearsed control, or cognitive distraction techniques.
[0060] The system also includes a classification module 110 configured to generate deception likelihood scores and corresponding confidence levels based on the fused outputs. The classification module 110 generates deception likelihood scores using probabilistic inference models, recurrent neural networks, or support vector machines trained on fused synchrony features.
[0061] The system also includes an explainable decision interface 112 configured to provide interpretable outputs highlighting contributing neural features and time windows relevant to deception classification. The explainable decision interface 112 provides natural language summaries to enable forensic, clinical, or legal experts to interpret the system’s outputs without requiring specialized knowledge of neural signal processing.
[0062] The system also includes a portable deployment unit 114 comprising hardware and software elements that enable real-time operation of the system for forensic, clinical, or security applications. The portable deployment unit 114 comprises wearable EEG headsets integrated with wireless communication modules, enabling real-time remote deception assessment. The portable deployment unit 114 further comprises cloud-based storage and computation facilities to support large-scale deployment across forensic and security networks.
[0063] FIG. 1 illustrates a flowchart outlining sequential step involved in a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models.
[0064] At 102, the system incorporates a multi-channel EEG data acquisition module, which is responsible for recording brain activity signals in real time as the subject responds to stimuli, questions, or interrogative prompts. This module ensures that the subtle neural fluctuations associated with cognitive load, hesitation, stress, or emotional arousal are captured in high resolution. The use of multiple EEG channels allows simultaneous monitoring of distributed brain regions, which is critical since deception involves complex interactions between frontal executive networks, limbic emotional centers, and parietal processing regions.
[0065] At 104, once the brain signals are collected, they are passed into the dual-branch deep learning processing unit, which has been specifically designed to separate and process two distinct aspects of neural functioning: cognition and emotion. The cognitive analysis branch examines neural patterns indicative of mental processes such as hesitation, increased workload, or executive conflict that arise when fabricating a lie or suppressing the truth. In parallel, the emotional analysis branch processes neural signals linked to affective responses such as stress, guilt, arousal, or anxiety that are typically activated during dishonest behavior. By structurally dividing the processing into two branches, the system ensures that both dimensions of deception cognitive and emotional are independently examined before being recombined for a more holistic interpretation.
[0066] At 106, after the dual-branch analysis, the signals are further refined using the spatiotemporal attention module. This component is designed to highlight the most relevant parts of the EEG data, both in terms of spatial regions of the brain and temporal windows in which significant changes occur. For example, not all brain regions contribute equally to deception, and not every time interval during a response is informative. The attention mechanism assigns dynamic weights to EEG features, amplifying patterns that are strongly indicative of deception while suppressing irrelevant or noisy signals. This ensures that the deep learning branches focus on the most critical features, thereby improving accuracy and robustness of the overall system.
[0067] At 108, the refined outputs from both the cognitive and emotional branches are then passed into the synchrony fusion module. This is the heart of the system’s novelty, as it integrates the two separate streams of information to analyze how cognition and emotion interact during deception. Rather than treating emotion and cognition as isolated indicators, the synchrony fusion module models their interplay. Deception often manifests as simultaneous activation of cognitive conflict and emotional arousal, and the synchrony of these signals creates unique neural patterns that cannot be detected by analyzing either pathway alone. The fusion module identifies these co-activations or disturbances in synchrony, which are strong indicators of dishonest behavior, even in cases where individuals attempt to suppress emotions or rehearse their responses.
[0068] At 110, the fused signals are directed into the classification module. This module applies trained algorithms to determine the likelihood that a given response is deceptive. It generates deception likelihood scores along with corresponding confidence levels, thus quantifying not just whether a response is likely deceptive but also how certain the system is of its decision. The classification step is critical in ensuring reliability and interpretability, especially in high-stakes scenarios such as forensic investigations or security screenings. Moreover, the system is robust to countermeasures since disturbances in synchrony patterns between cognitive and emotional signals reveal inconsistencies that betray suppression attempts.
[0069] At 112, to make the outcomes usable in real-world applications, the classified results are passed to an explainable decision interface. This component addresses one of the major limitations of traditional black-box artificial intelligence models. Instead of merely providing a deception label, the explainable interface highlights the contributing neural features, specific brain regions, and critical time windows that influenced the decision. For instance, it may indicate that a spike in frontal theta activity combined with limbic stress markers at a specific time interval strongly contributed to the classification of deception. Such interpretability not only builds trust in the system but also allows professionals in forensic, clinical, or security domains to better understand the neural underpinnings of the decision, thereby increasing transparency and accountability.
[0070] At 114, the entire system is integrated into a portable deployment unit. This unit packages the hardware and software elements into a form that can be readily used in real-world scenarios. The portable design ensures that the system can operate in varied environments such as interrogation rooms, courtrooms, clinics, or field security checkpoints without requiring large-scale laboratory infrastructure. The deployment unit supports real-time operation, enabling immediate feedback about deception likelihood during live questioning or assessments. The portability, combined with real-time capabilities and explainable outputs, makes the system highly practical for diverse applications ranging from legal investigations to clinical diagnostics and national security.
[0071] FIG. 2 illustrates a flowchart showing working of a multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models.
[0072] The process begins with EEG signal acquisition, where multi-channel brain signals are recorded from a subject in real time. This step is critical as it forms the raw input for the entire system, capturing neural oscillations and activity patterns associated with responses to stimuli or questions.
[0073] Once the EEG signals are collected, they undergo signal preprocessing. This stage involves cleaning and refining the raw EEG data by removing noise, artifacts, and irrelevant frequencies. Preprocessing ensures that the input signals are of sufficient quality and clarity to allow meaningful feature extraction. Common techniques such as filtering, artifact removal, normalization, and segmentation are applied here. The goal is to isolate clean, representative brain activity signals that can be mapped to either emotional or cognitive processes.
[0074] Following preprocessing, the system performs parallel feature extraction, where the EEG signals are split into two distinct data streams. One stream focuses on emotional features, capturing aspects such as stress, arousal, or guilt-related neural markers. The other stream emphasizes cognitive features, including workload, hesitation, and mental conflict. By creating parallel streams, the system simultaneously examines how the brain handles both affective and executive functions during responses. This separation reflects the theoretical framework that deception is not solely a cognitive phenomenon but also involves an interplay with emotional states.
[0075] The extracted features are then processed by dual deep learning branches, with each branch dedicated to one feature stream. The cognitive branch learns to recognize neural signatures of cognitive strain or conflict, while the emotional branch is trained to identify signals associated with emotional activation. Each branch independently processes its respective input, allowing the system to capture detailed patterns unique to both domains of mental functioning.
[0076] To further refine feature selection, each branch incorporates a spatiotemporal attention layer. This mechanism enables the model to focus selectively on the most informative brain regions and critical time intervals. Instead of treating all signals equally, the attention mechanism dynamically assigns weight to certain spatial or temporal components, highlighting the parts of the EEG data that are most relevant for detecting deception. This makes the system more adaptive and sensitive to subtle variations in neural activity.
[0077] The outputs of the dual branches are then integrated through an emotion-cognition synchrony fusion module. This component is central to the system’s novelty, as it models how cognitive and emotional processes interact when deception occurs. The synchrony fusion module captures correlations, alignments, or disturbances between emotional arousal and cognitive conflict. For example, a deceptive response may exhibit simultaneous activation of stress-related signals and hesitation-related cognitive markers, which together create a distinctive synchrony pattern.
[0078] Once the fused signals are generated, they are fed into the deception classification module. This module classifies responses as truthful or deceptive by analyzing the synchrony-informed patterns. It generates a deception likelihood score along with a confidence level. Importantly, this step also accounts for countermeasures, such as deliberate emotional suppression or rehearsed responses, by identifying anomalies in the synchrony dynamics.
[0079] Finally, the classified output is passed to the explainability module and output generation unit. This module ensures that the system is not a black box but instead provides interpretable insights. It highlights the neural features, brain regions, and time intervals that contributed most to the deception decision. The explainability component enhances the reliability and usability of the system, particularly in sensitive domains like legal investigations, clinical assessments, and security screenings. The generated output includes both the deception likelihood and an interpretable explanation that can guide practitioners in understanding the neural basis of the system’s decision.
[0080] While the invention has been described in connection with what is presently considered to be the most practical and various embodiments, it will be understood that the invention is not to be limited to the disclosed embodiments, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the scope of the appended claims.
[0081] A person of ordinary skill in the art may be aware that, in combination with the examples described in the embodiments disclosed in this specification, units and algorithm steps may be implemented by electronic hardware, computer software, or a combination thereof.
[0082] The foregoing descriptions of specific embodiments of the present disclosure have been presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the present disclosure to the precise forms disclosed, and many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described to best explain the principles of the present disclosure and its practical application, and to thereby enable others skilled in the art to best utilize the present disclosure and various embodiments with various modifications as are suited to the particular use contemplated. It is understood that various omissions and substitutions of equivalents are contemplated as circumstances may suggest or render expedient, but such omissions and substitutions are intended to cover the application or implementation without departing from the scope of the present disclosure.
[0083] Disjunctive language such as the phrase “at least one of X, Y, Z,” unless specifically stated otherwise, is otherwise understood with the context as used in general to present that an item, term, etc., may be either X, Y, or Z, or any combination thereof (e.g., X, Y, and/or Z). Thus, such disjunctive language is not generally intended to, and should not, imply that certain embodiments require at least one of X, at least one of Y, or at least one of Z to each be present.
[0084] In a case that no conflict occurs, the embodiments in the present disclosure and the features in the embodiments may be mutually combined. The foregoing descriptions are merely specific implementations of the present disclosure, but are not intended to limit the protection scope of the present disclosure. Any variation or replacement readily figured out by a person skilled in the art within the technical scope disclosed in the present disclosure shall fall within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be subject to the protection scope of the claims.
, Claims:I/We Claim:
1. A multi-task EEG-based neurobehavioral lie detection system using emotion-cognition synchrony and deep attention models (100) comprising:
a multi-channel EEG data acquisition module (102) configured to record brain activity signals associated with a subject’s responses in real-time;
a dual-branch deep learning processing unit (104) comprising a cognitive analysis branch and an emotional analysis branch;
a spatiotemporal attention module (106) configured to dynamically identify and emphasize EEG signal features from relevant brain regions and critical time intervals;
a synchrony fusion module (108) configured to integrate outputs from the cognitive analysis branch and the emotional analysis branch to capture interactions between cognitive and emotional responses during deceptive behavior;
a classification module (110) configured to generate deception likelihood scores and corresponding confidence levels based on the fused outputs;
an explainable decision interface (112) configured to provide interpretable outputs highlighting contributing neural features and time windows relevant to deception classification;
a portable deployment unit (114) comprising hardware and software elements that enable real-time operation of the system for forensic, clinical, or security applications.
2. The system (100) as claimed in claim 1, wherein the dual-branch deep learning processing unit (104) is trained using labeled datasets comprising truthful and deceptive responses, thereby enabling supervised learning of deception-related neural patterns.
3. The system (100) as claimed in claim 1, wherein the cognitive analysis branch of the dual-branch deep learning processing unit (104) is specifically configured to detect workload, hesitation, or mental conflict through analysis of frequency band activity within frontal and parietal EEG regions.
4. The system (100) as claimed in claim 1, wherein the emotional analysis branch of the dual-branch deep learning processing unit (104) is configured to identify stress, guilt, or arousal-related neural signatures by monitoring event-related potentials and power spectral densities.
5. The system (100) as claimed in claim 1, wherein the spatiotemporal attention module (106) applies weighted feature selection to dynamically prioritize temporal segments and spatial channels of EEG data most relevant for deception detection.
6. The system (100) as claimed in claim 1, wherein the classification module (110) generates deception likelihood scores using probabilistic inference models, recurrent neural networks, or support vector machines trained on fused synchrony features.
7. The system (100) as claimed in claim 1, wherein disturbances in synchrony patterns detected by the synchrony fusion module (108) are utilized to identify countermeasures including emotional suppression, rehearsed control, or cognitive distraction techniques.
8. The system (100) as claimed in claim 1, wherein the explainable decision interface (112) provides natural language summaries to enable forensic, clinical, or legal experts to interpret the system’s outputs without requiring specialized knowledge of neural signal processing.
9. The system (100) as claimed in claim 1, wherein the portable deployment unit (114) comprises wearable EEG headsets integrated with wireless communication modules, enabling real-time remote deception assessment.
10. The system (100) as claimed in claim 1, wherein the portable deployment unit (114) further comprises cloud-based storage and computation facilities to support large-scale deployment across forensic and security networks.