Search

CN-122000057-A - EEG-based pension institution visual environment cognition performance evaluation method

CN122000057ACN 122000057 ACN122000057 ACN 122000057ACN-122000057-A

Abstract

An EEG-based evaluation method for cognitive performance of vision environment of a pension institution relates to the technical field of environmental monitoring, and aims at solving the problem that the cognitive performance of a scheme cannot be effectively prejudged and optimized in the prior art, the cognitive task and the cognitive state classification model are combined, objective and quantitative evaluation of the cognitive performance of the vision environment of the pension institution is realized, the problem of low evaluation accuracy of the traditional design evaluation method is solved, and further, the cognitive performance of the scheme is effectively prejudged and optimized.

Inventors

  • WEI DAKE
  • Luo Kaizhou

Assignees

  • 哈尔滨工业大学

Dates

Publication Date
20260508
Application Date
20260126

Claims (10)

  1. 1. An EEG-based pension institution visual environment cognition performance evaluation method is characterized by comprising the following steps: step 1, acquiring visual environment images of a pension institution, and acquiring EEG signals generated when a tester observes the images; extracting multi-domain features of EEG signals to obtain brain electrical feature vectors, wherein the multi-domain features comprise time domain, frequency domain, time frequency and space features, and classifying the brain electrical feature vectors by using a cognitive task classification model and a cognitive performance classification model to obtain cognitive task labels and cognitive performance labels corresponding to vision environment images of a pension institution; step 3, inputting the cognitive task label and the cognitive performance label into an environmental cognitive performance evaluation model to obtain an output environmental cognitive performance index; and 4, obtaining an environmental cognition performance evaluation result according to the environmental cognition performance index.
  2. 2. An EEG-based pension facility visual environment cognitive performance assessment method according to claim 1, wherein said cognitive task classification model and cognitive performance classification model are trained by: Step 21, creating a psychological experiment paradigm for guiding the old to conduct cognitive ability test; step 22, acquiring EEG signals generated by the old people when the psychological experiment paradigm is executed, and generating cognitive task labels and cognitive performance labels corresponding to the EEG signals according to task events in the psychological experiment paradigm; Step 23, extracting multi-domain features of EEG signals to obtain brain electrical feature vectors; And step 24, training a cognitive task classification model by using the brain electrical feature vector and the corresponding cognitive task label, and training a cognitive performance classification model by using the brain electrical feature vector and the corresponding cognitive performance label.
  3. 3. An EEG-based pension facility visual environment cognitive performance assessment method according to claim 2, wherein said step 22 comprises the specific steps of: Step 221, in the psychology experiment model, each task stimulus presentation time is used as a key event, and the presentation event of each stimulus is recorded as an event mark; Step 222, selecting a corresponding time window based on the front-back time range of stimulus presentation according to each event mark, and extracting EEG signals in the corresponding time window, wherein EEG signal segments are epochs, and each epoch corresponds to an electroencephalogram response when stimulus presentation of a task event; Step 223, after the EEG signal is extracted and the cognitive tasks and cognitive performance tags are labeled, the cognitive performance tags are aligned with epoch.
  4. 4. An EEG-based pension institution visual environment cognitive performance assessment method according to claim 3, characterized in that said environment cognitive performance assessment model is obtained by: step 31, extracting environmental features of visual environmental images of the pension institutions as environmental cognitive performance indexes, wherein the environmental features comprise interface dimension features, space dimension features and semantic dimension features; step 32, acquiring a cognitive performance label and an environmental cognitive performance index of the same cognitive task standard sign as training data, respectively training a plurality of machine learning regression models, performing cross-validation on different regression models, and then obtaining an optimal parameter combination through a Bayesian optimization method; And step 33, estimating the prediction capacity and error range of the regression model through the mean square error, the root mean square error, the average absolute error and the decision coefficient, and selecting the regression model with optimal performance as an environmental cognition performance estimation model according to the estimation result.
  5. 5. An EEG-based pension facility visual environment cognition performance assessment method according to claim 1, characterised in that said interface dimensional features comprise colour complexity, shape complexity and texture complexity.
  6. 6. An EEG-based pension facility visual environment cognitive performance assessment method according to claim 5, wherein said colour complexity is expressed as: ; ; Wherein, the For the region of the partial mask, For the visual complexity of the partial mask, For the location of the locally masked pixels, The average color value within the partial mask, For the CIELab color value differences between each pixel pair, A is the standard deviation of the color difference in the partial mask for the gaussian weighting function, Average color value Expressed as: ; wherein N is the total number of pixels in the partial mask; ; Wherein, the As a result of the normalization factor, For Euclidean distance in the CIELab color space, expressed as: ; wherein L, a and b are respectively 3 channels of the image CIELab color space, 、 、 The values of the channels in the CIELab color space at the pixel locations within the region mask, 、 、 The values of the channels in the CIELab color space at the center of the area mask, respectively.
  7. 7. An EEG-based pension facility visual environment cognitive performance assessment method according to claim 6, wherein said shape complexity is expressed as: ; Wherein, the Is a characteristic binary group The probability of the occurrence of the presence of a defect, Is the gray value of the pixel, Is the neighborhood gray scale average value.
  8. 8. An EEG-based pension facility visual environment cognitive performance assessment method according to claim 7, characterised in that the texture complexity is expressed as: ; Wherein, the As a parameter of the dimensions of the device, Expressed at a given scale Lower, the minimum number of boxes required to be able to cover object O; The space dimension features are obtained through a ZoeDepth depth prediction model and comprise a space depth average value, a space depth standard deviation and a foreground, a middle scene, a background and a distant scene proportion of a visual environment; the spatial depth average calculation formula is as follows: ; Wherein, the S is the sum of the number of all pixel points for the depth value of each pixel point; the spatial depth standard deviation calculation formula is as follows: ; Wherein, the Is the spatial depth average; Foreground, middle, background and distant view proportions Expressed as: ; Wherein, the In order to represent the number of pixels occupied by 4 space types in a depth image corresponding to a visual scene, the depth values of a foreground, a middle scene, a background and a far background are respectively as follows: 、 、 And S is the sum of the number of all pixel points; The semantic dimension features are expressed as: ; Wherein, the Represents the proportion of each visual element in the image, p represents the pixel number occupied by a certain visual element, The visual element classification function is represented as such, Representing the total number of pixels of the image.
  9. 9. An EEG based pension facility visual environment cognition performance assessment method according to claim 1, characterised in that said step 1 further comprises the step of pre-processing the EEG signal, said pre-processing comprising channel selection, filtering, de-artefact, down-sampling and segmentation.
  10. 10. The EEG-based pension machine vision environment cognition performance assessment method according to claim 1, wherein said pension machine vision environment images comprise pension machine bedroom images, pension machine living room images, pension machine corridor images, pension machine restaurant images or pension machine hall images, and the camera height remains the same as the average height of the elderly when the images are acquired.

Description

EEG-based pension institution visual environment cognition performance evaluation method Technical Field The application relates to the technical field of environmental monitoring. In particular to an EEG-based assessment method for cognitive performance of vision environment of a pension institution. Background Elderly people often experience reduced cognitive function with age, significantly affecting their independence, and limiting the fundamental ability of daily life, a process known as cognitive decline. Cognitive decline manifests as a decline in core functions such as attention, memory, sensory function, and executive function, and may progress gradually to mild cognitive impairment and dysthymia. Because drug therapies have limited efficacy for nootropic symptoms and may be accompanied by adverse reactions, non-drug therapies (e.g., environmental interventions, physical activity and social participation, etc.) are considered important supportive and alternative interventions. The pension institution mainly provides pension service for elderly people such as senior, disabled and maldevelopment, and about 30% -65% of the incorporators suffer from maldevelopment. Environmental intervention is a key component of non-drug therapy, and is helpful to maintain and improve the self-care ability of the elderly by compensating for cognitive impairment through environmental support. Elderly people living in the aged for a long time are commonly exposed to indoor visual environments, and cognitive performance of the elderly people is mainly influenced by the indoor visual environments. Currently, cognition friendly environment design and optimization of a pension institution mainly depend on experience judgment of architects, caregivers and managers, and are assisted by subjective methods such as questionnaire investigation or behavior observation. The subjective factors cause the problem of low evaluation accuracy in the traditional design evaluation method, which results in the problem that the cognitive performance of the scheme cannot be effectively prejudged and optimized in the design stage of the indoor visual environment of the pension institution. Disclosure of Invention The invention aims to provide an EEG-based assessment method for cognitive performance of a vision environment of a pension institution, aiming at the problem that the cognitive performance of a scheme cannot be effectively prejudged and optimized in the prior art. The technical scheme adopted by the invention for solving the technical problems is as follows: An EEG-based pension institution visual environment cognitive performance assessment method comprises the following steps: step 1, acquiring visual environment images of a pension institution, and acquiring EEG signals generated when a tester observes the images; extracting multi-domain features of EEG signals to obtain brain electrical feature vectors, wherein the multi-domain features comprise time domain, frequency domain, time frequency and space features, and classifying the brain electrical feature vectors by using a cognitive task classification model and a cognitive performance classification model to obtain cognitive task labels and cognitive performance labels corresponding to vision environment images of a pension institution; step 3, inputting the cognitive task label and the cognitive performance label into an environmental cognitive performance evaluation model to obtain an output environmental cognitive performance index; and 4, obtaining an environmental cognition performance evaluation result according to the environmental cognition performance index. Further, the cognitive task classification model and the cognitive performance classification model are obtained through training of the following steps: Step 21, creating a psychological experiment paradigm for guiding the old to conduct cognitive ability test; step 22, acquiring EEG signals generated by the old people when the psychological experiment paradigm is executed, and generating cognitive task labels and cognitive performance labels corresponding to the EEG signals according to task events in the psychological experiment paradigm; Step 23, extracting multi-domain features of EEG signals to obtain brain electrical feature vectors; And step 24, training a cognitive task classification model by using the brain electrical feature vector and the corresponding cognitive task label, and training a cognitive performance classification model by using the brain electrical feature vector and the corresponding cognitive performance label. Further, the specific steps of the step 22 are as follows: Step 221, in the psychology experiment model, each task stimulus presentation time is used as a key event, and the presentation event of each stimulus is recorded as an event mark; Step 222, selecting a corresponding time window based on the front-back time range of stimulus presentation according to each event mark, and ext