Reading ability assessment method and system based on EEG signal and storage medium

Abstract

The invention relates to a reading ability evaluation method and system based on EEG signals and a storage medium. The method comprises the following steps: acquiring an EEG signal when a user reads in real time, and extracting specific neural feature indexes (including gamma frequency band power, theta-alpha phase synchronism, beta frequency band ERD and theta / beta ratio) corresponding to a cognitive sub-process; evaluating a reading understanding ability level based on the index; and dynamically adjusting the reading material content. According to the system, the evaluation robustness is improved through multi-modal data (EEG, eye movement and answer) fusion, and real-time feedback signals (such as prompt tone and white noise) are generated to intervene in attention distraction or cognize the overload state. According to the method, fine diagnosis and self-adaptive learning closed loop of the reading capability are realized, and the technical defects of lagging, coarse granularity and lack of targeted intervention of a traditional method are overcome.

Description

Technical Field

[0001] This invention relates to the fields of educational technology and artificial intelligence technology, and in particular to a reading ability assessment method, system, and storage medium based on EEG signals. Background Technology

[0002] Reading comprehension is a core cognitive ability for humans to acquire information and learn knowledge, and it is crucial for children's academic achievement and future development. Therefore, scientifically and accurately assessing and improving children's reading comprehension ability has always been a research focus in fields such as education, psychology, and cognitive neuroscience.

[0003] Traditional reading ability assessments rely on standardized tests, which suffer from problems such as time lag, coarse-grained results, and lack of ecological validity. While eye-tracking technology can monitor behavior in real time, it cannot accurately distinguish cognitive subprocesses (such as decoding, integration, and reasoning). Existing EEG technologies mostly focus on generalized cognitive load assessment, lacking indicators for specific neural mechanisms of reading, and are not integrated with adaptive learning loops, making it difficult to achieve personalized intervention. Summary of the Invention

[0004] In view of the above situation, it is necessary to provide a reading ability assessment method based on EEG signals to solve at least one of the above problems, characterized by including the following steps:

[0005] While the user is performing a reading task, the user's EEG signal is acquired;

[0006] From the EEG signals, at least one specific neural feature index corresponding to a specific cognitive process of reading comprehension is extracted; the specific neural feature index includes: gamma band power related to the word decoding process, phase synchronization of the theta and alpha bands related to the semantic integration process, and event-related desynchronization (ERD) of the beta band related to the reasoning process.

[0007] Based on the at least one specific neural characteristic indicator, the user's current reading comprehension ability level is assessed and determined.

[0008] The content of the reading materials for the reading task is dynamically adjusted based on the reading comprehension ability level.

[0009] Preferably, the method further includes extracting the θ / β ratio related to the attention state and using the θ / β ratio to comprehensively assess the user's current reading comprehension ability level.

[0010] Preferably, the step of assessing and determining the user's current reading comprehension ability level based on the at least one specific neural feature index specifically includes:

[0011] The extracted values ​​of the specific neural feature indicators are compared with a preset evaluation model; the evaluation model presets the value ranges of the specific neural feature indicators corresponding to different reading comprehension ability levels.

[0012] Preferably, the step of dynamically adjusting the content of the reading material according to the reading comprehension ability level specifically includes:

[0013] Based on the assessed reading comprehension level, the reading material should be adjusted by selecting at least one dimension from four dimensions: vocabulary complexity, sentence structure complexity, text coherence, and question type.

[0014] Preferably, when the user is performing a reading task, their eye movement data and answer performance data are acquired simultaneously;

[0015] A Bayesian fusion algorithm is used to combine the specific neural feature indicators, the eye movement data, and the answer performance data to jointly assess and determine the user's current reading comprehension ability level.

[0016] Preferably, a real-time feedback signal is generated based on the specific neural characteristic indicators;

[0017] When a decrease in the β-band ERD corresponding to excessive cognitive load is detected, a first preset feedback signal is output.

[0018] And / or, when an increase in the θ / β ratio corresponding to attentional distraction is detected, a second preset feedback signal is output.

[0019] Preferably, the first preset feedback signal is a prompt tone or a pause suggestion, and the second preset feedback signal is white noise.

[0020] This application also proposes a scheme for a reading ability assessment system based on EEG signals, comprising:

[0021] The signal acquisition module is used to acquire the user's EEG signal when the user is performing a reading task;

[0022] The processing module, connected to the signal acquisition module, is used to: extract at least one specific neural feature indicator corresponding to a specific cognitive process of reading comprehension from the EEG signal, wherein the specific neural feature indicator includes: gamma band power related to the word decoding process, phase synchronization of the theta and alpha bands related to the semantic integration process, and beta band event-related desynchronization (ERD) related to the reasoning process; assess and determine the user's current reading comprehension ability level based on the at least one specific neural feature indicator; and generate adjustment instructions for adjusting the reading material content of the reading task.

[0023] The application module, connected to the processing module, is used to dynamically adjust the reading material content presented to the user according to the adjustment instructions.

[0024] Preferably, the signal acquisition module is further used to acquire the user's eye movement data and answer performance data;

[0025] The processing module is also used to employ a Bayesian fusion algorithm to combine the specific neural feature indicators, the eye movement data, and the answer performance data to jointly assess and determine the user's current reading comprehension ability level.

[0026] In this application, a solution is also proposed: a computer-readable storage medium having a computer program stored thereon, characterized in that the program, when executed by a processor, implements the method described in any of the above claims.

[0027] (III) Beneficial Effects

[0028] The depth and accuracy of the assessment are significantly improved: This invention creatively maps specific EEG neural features one-to-one with the core cognitive sub-processes of reading comprehension (decoding, integration, and reasoning). This allows the assessment to move beyond the general level of "cognitive load" and, like a "diagnostic scan," precisely pinpoint the specific weaknesses in the user's reading cognitive chain. This diagnostic depth is unparalleled by existing technologies.

[0029] This invention achieves true real-time dynamic intervention: It boasts a response time as short as 10 milliseconds. Once neural signals of distraction or cognitive overload are detected, the system can immediately trigger audiovisual feedback for immediate intervention, effectively guiding the user's cognitive state back on track, preventing error accumulation, and greatly improving the efficiency and experience of the learning process.

[0030] This invention constructs a complete closed loop for personalized learning: seamlessly integrating "precise assessment" with "adaptive content adjustment." Based on real-time diagnosed skill gaps (such as weak vocabulary decoding), the system intelligently pushes targeted training content (such as simplified vocabulary), achieving dynamic scaffolded teaching within the user's "zone of proximal development." This forms a virtuous cycle of "assessment-diagnosis-intervention-improvement," realizing truly one-on-one, personalized tutoring.

[0031] The robustness and reliability of the evaluation results are stronger: By integrating data from three different modalities—EEG, eye tracking, and behavioral responses—this invention utilizes cross-validation of multi-source information to effectively overcome the risk of misjudgment that may arise from a single data source (such as noise in EEG signals or ambiguity in eye tracking data), making the final evaluation results more stable and reliable, and providing solid data support for personalized education decisions. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the overall architecture of a reading ability assessment system provided in an embodiment of the present invention.

[0033] Figure 2 This is a schematic diagram showing the connection relationship of the main hardware layer components in an embodiment of the present invention.

[0034] Figure 3 This is a flowchart of a reading ability assessment method provided in an embodiment of the present invention.

[0035] Figure 4 This is a schematic diagram of the EEG signal processing and feature extraction sub-process in an embodiment of the present invention.

[0036] Figure 5 This is a schematic diagram of the decision-making process of the adaptive content adjustment logic in an embodiment of the present invention.

[0037] Figure 6 This is a schematic diagram of the multimodal data fusion evaluation model in an embodiment of the present invention.

[0038] Figure 7 This is an example diagram of a real-time feedback user interface in an embodiment of the present invention, demonstrating a visualization of cognitive state in the form of a radar chart.

[0039] Figure 8 This is the evaluation criteria table in the embodiments of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the following detailed description of the reading ability assessment method, system, and storage medium based on EEG signals, in conjunction with the accompanying drawings and embodiments, is provided. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention.

[0041] In the description of this invention, unless otherwise stated, "a plurality of" means two or more; the terms "center," "longitudinal," "lateral," "upper," "lower," "left," "right," "inner," "outer," "front end," "rear end," "head," "tail," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0042] In the description of this utility model, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will be able to understand the specific meaning of the above terms in this utility model based on the specific circumstances.

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0044] Example

[0045] This embodiment provides a complete system for assessing and training children's reading ability based on EEG signals.

[0046] System Composition (Reference) Figure 1 and Figure 2 )

[0047] like Figure 1 As shown, the system is divided into a hardware layer, a data processing layer, and an application layer in terms of architecture.

[0048] Hardware layer (100): Responsible for the collection of raw data and the presentation of human-computer interaction content.

[0049] Data processing layer (200): This is the core of the system, responsible for signal processing, feature extraction, capability assessment and decision generation.

[0050] Application layer (300): Responsible for providing users with specific reading content, real-time feedback, and adaptive adjustments.

[0051] like Figure 2 As shown, the hardware layer (100) specifically includes:

[0052] EEG acquisition device (110): A portable 8-channel dry electrode EEG headband is used. Dry electrodes are chosen for easy and quick wear by children, eliminating the need for conductive gel. The 8 electrodes are arranged in brain regions closely related to language, attention, and integration functions according to the international 10-20 standard system, specifically: T5, T6, T3, and T4 in the temporal lobe (near Wernicke's area and Broca's area, responsible for language comprehension and production), P3 and P4 in the parietal lobe (responsible for multi-sensory information integration and attention allocation), Fz in the frontal lobe (as a reference or for monitoring frontal lobe activity), and Oz in the occipital lobe (monitoring visual input). The device has a sampling rate of 500Hz, and the built-in module performs preliminary signal amplification and analog-to-digital conversion. It connects to the data processing unit (130) via Bluetooth 5.0 or USB-C interface.

[0053] Eye-tracking device (120): A near-infrared eye tracker integrated below the display device, with a sampling rate of not less than 90Hz, capable of accurately capturing data such as the user's gaze point, pupil diameter, saccades and saccade paths.

[0054] Data Processing Unit (130): Employs the NVIDIA Jetson Nano Developer Kit as its edge computing core. Its quad-core ARM Cortex-A57 CPU and 128-core Maxwell GPU provide ample computing power for parallel processing of EEG signals and running machine learning models. The unit includes 64GB of eMMC storage and 4GB of LPDDR4 memory.

[0055] Display device (140): It features a 10.1-inch main screen (2560x1600 resolution, 120Hz refresh rate) for displaying reading text, and can also be equipped with a 5.5-inch secondary screen for displaying real-time feedback radar charts or interactive questions.

[0056] Interactive device (150): including keyboard and mouse, for users to answer questions or perform other operations.

[0057] The functions of the data processing layer (200) and the application layer (300) are implemented by a software system running on the data processing unit (130).

[0058] Method and Flow (Reference) Figure 3 )

[0059] The reading ability assessment method of the present invention has the following detailed process: Figure 3 As shown, it includes the following steps:

[0060] Step S301: Data Synchronization Acquisition

[0061] The user wears the EEG acquisition device (110) and faces the display device (140). The system begins an evaluation session. Throughout the user's reading task (such as reading an article and answering questions), the system synchronously and continuously records the following three types of data through the signal acquisition module:

[0062] Raw EEG signal from multiple channels (8 channels).

[0063] High-frequency eye-tracking data streams (fixation point coordinates, pupil size, etc.).

[0064] User behavior event markers, such as the time when reading begins, the time when turning pages, the time when a question is answered, and the answer (correct / incorrect).

[0065] All data streams are accompanied by precise timestamps to ensure subsequent cross-modal time alignment and event-related analysis.

[0066] Step S302: Signal preprocessing and feature extraction

[0067] This step is one of the core technologies of this invention, and its sub-processes are as follows: Figure 4 As shown, the raw EEG signal needs to undergo a series of processing steps to extract meaningful neural features.

[0068] Sub-step S401: Preprocessing

[0069] The raw EEG signal is mixed with various noises and artifacts. The purpose of preprocessing is to "clean up" the signal.

[0070] Filtering: A bandpass filter from 1 Hz to 45 Hz is applied to the continuous EEG data. The lower limit of 1 Hz effectively filters out the slow baseline drift (DC drift) caused by minute head movements and electrode polarization, while the upper limit of 45 Hz filters out most high-frequency electromuscular (EMG) noise and environmental electromagnetic interference. Additionally, a 50 Hz notch filter is applied to eliminate power frequency interference.

[0071] Artifact Removal: Independent Component Analysis (ICA) is employed. ICA decomposes multi-channel EEG signals into a series of statistically independent source components. By analyzing the time-domain waveform, frequency-domain characteristics, and spatial topography of each component, artifact components related to eye movement (EOG), blinking, electrocardiogram (ECG), and muscle activity (EMG) can be identified. For example, components related to blinking typically exhibit sharp waveforms with high amplitude in the frontal lobe region. After identifying these artifact components, they are removed from the original signal, and the remaining "clean" components are reconstructed back into the EEG signal. To improve the automation of the identification process, eye-tracking data can be incorporated; when a blink event is detected, ICA components within the same time period are automatically labeled as artifact candidates.

[0072] Sub-step S402: Segmentation and Baseline Correction

[0073] The preprocessed continuous EEG data is divided into "epochs" based on behavioral event markers. For example, the moment a user begins reading each text segment is set as 0, a 200-millisecond period is extracted before this point as the baseline period, and the entire period after the end of reading that segment is extracted as the task period. The baseline period represents the user's resting state without a specific cognitive task and will serve as the reference standard for subsequent calculations of features such as ERD.

[0074] Sub-step S403: Extraction of specific neural features

[0075] Based on the segmented data, four neural feature indicators related to specific cognitive subprocesses of reading comprehension were calculated:

[0076] Gamma Power (30-40Hz) - Characterization Terminology Decoding:

[0077] Calculation: For each task time period's EEG data segment, the time spectrum is calculated using Short Time Fourier Transform (STFT) or Wavelet Transform. Then, the average power spectral density within the 30-40Hz frequency band is extracted.

[0078] Cognitive Significance: Oscillatory activity in the gamma band is thought to be related to the synchronization of different neuronal clusters in the brain, forming the neural basis of feature binding. In reading, recognizing a word requires binding its visual features (graph shape), phonological features (pronunciation), and semantic features (meaning) together; this process is called lexical decoding or lexical access. Higher gamma band power, especially in the left temporoparietal region (e.g., T5, P3 electrodes), generally indicates a smoother and more efficient lexical decoding process.

[0079] Theta-Alpha Phase Synchrony - Characterization Semantic Integration:

[0080] Calculation: First, the EEG signal is filtered separately to obtain theta wave (4-8Hz) and alpha wave (8-13Hz) signals. Then, the phase synchronization between electrode pairs in different brain regions is calculated. For example, the theta-alpha cross-band phase synchronization index between frontal lobe electrodes (e.g., Fz) and posterior parietal-temporal lobe electrodes (e.g., P3, T5) is calculated. Commonly used algorithms include Phase Locking Value (PLV) or Weighted Phase Lag Index (wPLI). wPLI can better eliminate spurious synchronization caused by volume conduction effects.

[0081] Cognitive Significance: Reading comprehension is not merely about recognizing individual words, but more importantly, about stringing words together into grammatically correct sentences and understanding their overall meaning—a process known as semantic integration. Neuroscience research indicates that long-range neural connections, particularly those involving the frontal lobe (responsible for working memory and executive control) and the posterior brain regions (responsible for information storage and processing), are crucial for sentence-level integration. Phase synchronization in the theta-alpha band is considered the carrier of this long-range communication; stronger synchronization indicates greater cognitive resource investment and higher efficiency in the brain's integration of sentence structure and construction of contextual models.

[0082] Beta band event-related desynchronization (Beta ERD, 13-30Hz) - representation inference processing:

[0083] Calculation: Calculate the power of the β band (13-30Hz) during the task period and compare it with the β band power during the baseline period. The calculation formula is: ERD% = ((P_baseline - P_task) / P_baseline) * 100%, where P_baseline is the baseline power and P_task is the task power. ERD represents a decrease in power and is therefore a positive value.

[0084] Cognitive Significance: Desynchronization (i.e., power reduction) in the beta band is typically associated with cortical activation and the preparation and execution of sensorimotor or cognitive processes. During reading, when users need to reason beyond the surface information of the text, predict, or resolve ambiguities, they require the engagement of higher-level cognitive functions. At this time, the beta band in relevant brain regions (such as the parietal and frontal lobes) exhibits significant ERD (Early Desynchronization), reflecting a shift in the neuronal population from a synchronous resting state to an asynchronous, active computational state. Therefore, the degree of beta ERD can serve as an effective indicator of whether a user is engaging in deep reasoning processing.

[0085] Theta / Beta ratio – a measure of attentional state:

[0086] Calculation: Calculate the power in the θ band (4-8Hz) and β band (13-30Hz) of the frontal lobe region (e.g., Fz electrode), and then calculate the ratio between the two.

[0087] Cognitive Significance: The theta / β ratio is a well-established indicator in cognitive neuroscience used to measure attentional states. Generally, a higher theta / β ratio is associated with inattention, mind wandering, or drowsiness; while a lower theta / β ratio is associated with focused attention and engagement in the cognitive task. In reading tasks, real-time monitoring of this ratio can effectively determine whether a user has maintained focus on the reading material.

[0088] Step S303: Multimodal fusion evaluation

[0089] To obtain a comprehensive and robust evaluation result, this embodiment employs a multimodal data fusion strategy, such as... Figure 6 As shown.

[0090] Input feature vector: For each evaluation unit (such as a paragraph or a question), construct a multimodal feature vector, including:

[0091] EEG features (R): {γ_power,θ-α_sync,β_ERD,θ / β_ratio}

[0092] Eye movement characteristics (E): {mean fixation duration, number of fixations, number of regressions, mean pupil diameter}

[0093] Answer performance characteristics (A): {Answer accuracy rate, answer time}

[0094] Fusion Model: Employs a Naive Bayes classifier. The model aims to calculate the posterior probability P(C|R,E,A) of a user's reading comprehension ability belonging to a certain level C (e.g., "Excellent", "Good", "Average", "Poor") given observed features (R,E,A). According to Bayes' theorem: P(C|R,E,A)∝P(R|C)*P(E|C)*P(A|C)*

[0095] P(C)

[0096] Under the "naive" assumption, all features are conditionally independent given an ability level C. P(C) is the prior probability of the ability level. Conditional probabilities (likelihoods) such as P(R|C), P(E|C), and P(A|C) can be learned from a pre-labeled training dataset. For example, P(γ_power|C = "Excellent") can be estimated by statistically analyzing the distribution (e.g., Gaussian distribution) of γ power values ​​among all samples labeled "Excellent" by experts.

[0097] Level Determination: After calculating the posterior probability of each ability level, the level with the highest posterior probability is selected as the current user's reading comprehension ability level. In addition, the system will also refer to... Figure 8 The evaluation criteria table shown details the specific numerical ranges of each indicator corresponding to different levels, serving as a supplementary decision-making basis and a source of interpretability for the model.

[0098] Step S304: Real-time feedback and dynamic intervention

[0099] Based on the extracted neural features and the assessed ability level, the system performs two types of interventions:

[0100] Real-time feedback (for reference) Figure 7 ):

[0101] Visual feedback: A real-time updated "cognitive radar chart" is displayed on the secondary screen. Figure 7 The radar chart's four axes correspond to gamma power, θ-α synchronicity, βERD, and (inversely) the θ / β ratio, respectively. The dynamic polygon formed by the user's real-time metrics allows them to intuitively see their cognitive resource allocation across the four dimensions of decoding, integration, reasoning, and attention, thereby enhancing their metacognitive abilities.

[0102] Auditory feedback: When the βERD value is detected to be consistently below a preset threshold (e.g., a 50% decrease compared to the user's personal average), it indicates that the user may have encountered incomprehensible content and is experiencing excessive cognitive load. The system will play a gentle prompt tone and display a suggestion on the screen: "Do you need to take a break or check the prompt?"

[0103] When the θ / β ratio is consistently above a preset threshold (e.g., more than twice the average for a user under focused conditions), it indicates that the user may be distracted. The system will then play a low-volume white noise through the headphones, using the "random resonance" effect of auditory stimulation to gently redirect the user's attention back to the reading task.

[0104] Step S305: Adaptive Content Adjustment (Reference) Figure 5 )

[0105] This is a crucial step in forming a learning loop. Based on the ability level assessed in S303 and combined with the specific neural features extracted in S302, the system performs personalized content adjustments.

[0106] Decision logic (e.g.) Figure 5 ):

[0107] If the rating is "Excellent": The system will recommend more challenging "extended" reading materials and set reading goals that focus on deeper understanding and critical thinking (e.g., analyzing the author's implied meaning).

[0108] If the rating is "Good": The system will push "consolidation" materials of moderate difficulty, and set reading goals of summarizing and simple reasoning.

[0109] If the level is "normal": the system will push "basic" materials, along with some auxiliary questions to help extract key information.

[0110] If the rating is "poor": the system will further analyze the specific shortcomings.

[0111] If the main obstacle is low gamma power, it indicates that vocabulary decoding is the primary hurdle. The system will then push "simplified" text, such as replacing rare words with more common ones, or automatically annotating rare words with pinyin and definitions.

[0112] If the main issue is weak θ-α synchronicity, it indicates difficulty in sentence integration. The system will then push texts with simpler sentence structures, such as breaking down long or complex sentences into shorter or simpler ones.

[0113] If the βERD is not obvious, it indicates a lack of reasoning activity. The system will design more guiding reasoning questions after the text, gradually transitioning from factual questions to reasoning questions, training the reasoning ability in a "scaffolding" manner.

[0114] Step S306: Learning the closed loop and model iteration

[0115] All data generated by a user during a session (EEG, eye tracking, behavior, system intervention, ability changes) is recorded to form a personal learning profile. This data can be used to track the user's long-term progress and to periodically and personally update and optimize the multimodal fusion assessment model, making its assessment of the user increasingly accurate.

[0116] Through the above steps, the method and system disclosed in this invention successfully combine cutting-edge cognitive neuroscience findings with artificial intelligence education technology, breaking through the limitations of traditional reading assessment, realizing in-depth diagnosis and real-time personalized intervention of reading comprehension ability, and providing a new, efficient, and scientific paradigm for promoting the development of children's reading ability.

[0117] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A reading ability assessment method based on EEG signals, characterized in that, Includes the following steps: While the user is performing a reading task, the user's EEG signal is acquired; From the EEG signals, at least one specific neural feature indicator corresponding to a specific cognitive process of reading comprehension is extracted; The specific neural feature indicators include: γ-band power related to the lexical decoding process, phase synchronization of the θ-band and α-band related to the semantic integration process, and β-band event-related desynchronization (ERD) related to the reasoning process. Based on the at least one specific neural characteristic indicator, the user's current reading comprehension ability level is assessed and determined. The content of the reading materials for the reading task is dynamically adjusted based on the reading comprehension ability level.

2. The method according to claim 1, characterized in that, It also includes extracting the θ / β ratio related to attention state and using the θ / β ratio to comprehensively assess the user's current reading comprehension ability level.

3. The method according to claim 1 or 2, characterized in that, The step of determining the user's current reading comprehension ability level based on the at least one specific neural feature index specifically includes: The extracted values ​​of the specific neural feature indicators are compared with a preset evaluation model; the evaluation model presets the value ranges of the specific neural feature indicators corresponding to different reading comprehension ability levels.

4. The method according to claim 1, characterized in that, The step of dynamically adjusting the content of reading materials according to the reading comprehension ability level specifically includes: Based on the assessed reading comprehension level, the reading material should be adjusted by selecting at least one dimension from four dimensions: vocabulary complexity, sentence structure complexity, text coherence, and question type.

5. The method according to claim 1, characterized in that, Also includes: While the user is performing a reading task, their eye movement data and answer performance data are acquired simultaneously. A Bayesian fusion algorithm is used to combine the specific neural feature indicators, the eye movement data, and the answer performance data to jointly assess and determine the user's current reading comprehension ability level.

6. The method according to claim 1, characterized in that, Also includes: Based on the specific neural characteristic indicators, a real-time feedback signal is generated; When a decrease in the β-band ERD corresponding to excessive cognitive load is detected, a first preset feedback signal is output. And / or, when an increase in the θ / β ratio corresponding to attentional distraction is detected, a second preset feedback signal is output.

7. The method according to claim 6, characterized in that, The first preset feedback signal is a prompt tone or a pause suggestion, and the second preset feedback signal is white noise.

8. A reading ability assessment system based on EEG signals, characterized in that, include: The signal acquisition module is used to acquire the user's EEG signal when the user is performing a reading task; The processing module, connected to the signal acquisition module, is used to: extract at least one specific neural feature indicator corresponding to a specific cognitive process of reading comprehension from the EEG signal, wherein the specific neural feature indicator includes: gamma band power related to the word decoding process, phase synchronization of the theta and alpha bands related to the semantic integration process, and beta band event-related desynchronization (ERD) related to the reasoning process; assess and determine the user's current reading comprehension ability level based on the at least one specific neural feature indicator; and generate adjustment instructions for adjusting the reading material content of the reading task. The application module, connected to the processing module, is used to dynamically adjust the reading material content presented to the user according to the adjustment instructions.

9. The system according to claim 8, characterized in that, The signal acquisition module is also used to acquire the user's eye movement data and answer performance data; The processing module is also used to employ a Bayesian fusion algorithm to combine the specific neural feature indicators, the eye movement data, and the answer performance data to jointly assess and determine the user's current reading comprehension ability level.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 7.

Images