Multi-modal physiological signal fusion olfaction function evaluation device and evaluation method

Abstract

The invention belongs to the technical field of olfactory diagnosis devices, and particularly relates to an olfactory function evaluation device and method based on multi-modal physiological signal fusion, and the evaluation device comprises an olfactory stimulation module, a multi-modal physiological signal acquisition module, a synchronous control module and a data processing and analysis module; according to the olfactory function evaluation system, comprehensive evaluation of the olfactory function is achieved by synchronously collecting multi-mode physiological signals such as respiration, heart rate variability, myoelectricity, facial expression and eyeball movement and combining olfactory stimulation. According to the method, the limitation of single-mode evaluation is overcome, the accuracy and objectivity of olfactory function evaluation are improved, and the method has important clinical application value.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of olfactory diagnosis devices, and particularly relates to a multi-modal physiological signal fusion olfactory function evaluation device and evaluation method. BACKGROUND

[0002] Olfaction is one of the oldest senses of mammals, and is an important way for animals to perceive and adapt to the environment. Olfactory disorders can affect people's quality of life, social interaction, nutrient intake, and even threaten life safety, especially for the elderly population, the impact is more significant. In addition, olfactory disorders are also related to emotional disorders, mental diseases and neurodegenerative diseases such as Parkinson's disease and Alzheimer's disease.

[0003] Olfactory disorders refer to abnormal perception of odor caused by organic and / or functional lesions of each link in the olfactory pathway during the process of odor perception, transmission and information analysis and integration, including quantitative and qualitative disorders of olfaction. The former includes hyposmia (reduced sensitivity to odor perception), anosmia (inability to perceive odor) and hyperosmia (abnormal sensitivity to one or more odors), and the latter includes parosmia (distortion of odor property perception) and phantosmia (odor perception without any odor).

[0004] Currently, the diagnosis of olfactory dysfunction mainly relies on two methods of subjective psychophysical test evaluation and objective evaluation of olfaction, the former mainly includes T&T olfactometer test, Sniffin' Sticks olfactory stick test, University of Pennsylvania Smell Identification Test (UPSIT), and Connecticut Chemical Sensory Clinical Research Center Connecticut Olfactory Function Test, etc. Traditional objective evaluation methods of olfactory function include event-related potentials (ERPs), including olfactory event-related potentials (oERPs) and trigeminal event-related potentials (tERPS); imaging examinations include thin-slice CT and MRI of nasal cavity and paranasal sinuses, and magnetic resonance examination of olfactory pathway to evaluate the morphology of structures such as olfactory bulb, olfactory filament and olfactory tract. Subjective evaluation of olfaction is useful in the evaluation of most adults, but requires good compliance and active cooperation of patients or test subjects, and is easily affected by factors such as attention, cognitive state and memory ability of the subjects. Due to the influence of semantic factors related to odor naming, the application on children, patients with cognitive impairment, patients with neurodegenerative diseases and patients with different language / cultural backgrounds has certain limitations. Sniff Magnitude Test can simply, quickly and efficiently detect the advantages of the elderly, children and others because it is objective, does not depend on language, cognition, memory, naming and other abilities, and does not require special subjective cooperation. It has been applied to objective testing of olfactory function for more than a decade. However, due to the limited accuracy of previous products and the fewer testable indicators, it has not been widely recognized. Although the objective evaluation method has high accuracy, it is technically complex, time-consuming and costly, and is only used in highly specialized medical centers on a non-routine basis, making it difficult to popularize in primary medical institutions. The correlation between subjective test and objective evaluation of olfactory function is poor, and more importantly, the existing olfactory function evaluation technology generally has the problem of single mode, such as the inability to capture brain electrical oscillations triggered by olfactory stimulation when only analyzing respiratory signals; and the lack of correlation of physiological responses such as respiratory peak frequency when relying solely on electroencephalogram signals; and the difficulty in reflecting the complete neural circuit of olfactory cognitive processing without integrating the coordinated changes of electromyogram signals and facial expressions. The existing equipment has not realized the spatio-temporal synchronous acquisition and intelligent fusion of multi-modal signals, making it difficult to achieve comprehensive and accurate evaluation of olfactory function. SUMMARY

[0005] Therefore, the present application provides a multi-modal physiological signal fusion olfactory function evaluation device, which adopts a multi-modal objective olfactory function evaluation device, and constructs an olfactory function evaluation system integrating real-time acquisition of multi-modal physiological signals, intelligent feature extraction and deep fusion analysis, which can objectively and accurately evaluate the olfactory function state of patients.

[0006] In order to realize the above technical scheme, the specific technical scheme adopted by the present application is as follows:

[0007] The application discloses a multi-modal physiological signal fusion olfactory function evaluation device, which comprises an olfactory stimulation module, a multi-modal physiological signal acquisition module, a synchronous control module and a data processing and analysis module.

[0008] The olfactory stimulation module acts on a subject to provide olfactory stimulus.

[0009] The synchronous control module realizes time synchronization of the olfactory stimulation module and the multi-modal physiological signal acquisition module.

[0010] The data processing and analysis module performs fusion analysis on the time-synchronized multi-modal features to establish an olfactory function evaluation model; and the evaluation model is used to realize quantitative evaluation of the olfactory function.

[0011] The multi-modal physiological signal acquisition module comprises:

[0012] The heart rate acquisition unit acts on the subject to acquire heart rate variability of the subject.

[0013] The electromyography acquisition unit acts on the subject to acquire electromyography signals of the trapezius muscle of the subject.

[0014] Further, the olfactory stimulation module comprises an olfactory agent bin, an air flow control system and a nasal catheter.

[0015] The olfactory agent bin stores olfactory stimulus; the olfactory stimulus comprises water and two kinds of pleasant gas and one kind of unpleasant gas which do not activate the trigeminal nerve but only activate the olfactory nerve.

[0016] The air flow control system comprises an air compressor and an air flow converter; the air compressor is connected with the air flow converter through an air pipe, the air flow converter is connected with the olfactory agent bin through an air pipe, the air flow converter is provided with a first air outlet and a second air outlet, the first air outlet outputs air, the second air outlet outputs air containing the olfactory stimulus, and the first air outlet and the second air outlet are connected with the nasal catheter through an air pipe.

[0017] Further, the multi-modal physiological signal acquisition module further comprises a facial expression and eyeball movement acquisition unit; the facial expression and eyeball movement acquisition unit comprises a camera; the camera is arranged on the front of the nasal catheter or an independent support and is used to acquire facial expression changes and eyeball movement trajectories of the subject.

[0018] Further, the synchronous control module comprises a clock synchronization unit, the clock of the olfactory stimulation module and each physiological signal acquisition unit in the multi-modal physiological signal acquisition module are synchronized through an RS485 bus protocol, a synchronization time point is set as a t0 time point, a breathing frequency detection period Th is taken as a guide period from the t0 time point, and the olfactory stimulation module and each physiological signal acquisition unit are coordinated to work.

[0019] Further, the data processing and analysis module comprises a signal preprocessing unit, a feature extraction unit and a fusion evaluation unit;

[0020] The signal preprocessing unit filters and denoises the collected multi-modal physiological signals;

[0021] The feature extraction unit extracts respiratory features, heart rate variability, electromyographic features, facial expression features and eye movement features from the preprocessed signals;

[0022] The fusion evaluation unit uses a machine learning algorithm or a deep learning model to analyze the multi-modal features collected by the feature extraction unit, and establishes the olfactory function evaluation model.

[0023] Further, the respiratory features include respiratory rate, tidal volume and peak inspiratory flow rate;

[0024] The heart rate variability includes heart rate and LF / HF ratio;

[0025] The electromyographic features include electromyographic amplitude and electromyographic frequency characteristics;

[0026] The facial expression features include the movement characteristics of the zygomatic major muscle and the corrugator muscle;

[0027] The eye movement features include pupil diameter change and gaze time.

[0028] Further, the olfactory function evaluation model is used to perform quantitative evaluation on the olfactory function; the results of quantitative evaluation include normal olfaction, reduced olfaction and loss of olfaction.

[0029] Meanwhile, the present application also proposes a multi-modal physiological signal fusion olfactory function evaluation method realized by the above-mentioned multi-modal physiological signal fusion olfactory function evaluation device, comprising the following steps:

[0030] S101: Collect the respiratory multi-modal feature information of multiple normal olfaction samples and multiple olfactory loss patient samples respectively in the presence and absence of an olfactory agent;

[0031] S101: Collect the respiratory multi-modal feature information of multiple normal olfaction samples and multiple olfactory loss patient samples respectively in the presence and absence of an olfactory agent;

[0032] S102: Analyze the multi-modal physiological signal data of each sample, and obtain the time sequence waveform diagram of each modal feature under the condition of the presence and absence of an olfactory agent after performing Fourier transform on the data; obtain the frequency, amplitude and peak value of different modal features according to the waveform diagram;

[0033] S103: Analyze the frequency characteristics, amplitude characteristics and peak value characteristics in the following aspects:

[0034] Difference value in the same waveform

[0035] Difference value of two waveform graphs of the same modal characteristics and different olfactory states

[0036] Data labeling is performed to obtain training samples;

[0037] S104: training an olfactory function evaluation model based on the training samples;

[0038] S105: evaluating a non-quantitative subject based on the trained olfactory function evaluation model to determine the olfactory state of the non-quantitative subject.

[0039] Further, in the S102, when drawing the time series waveform graph of the breathing characteristics:

[0040] If the start time of releasing the odorant is at the amplitude peak of the waveform graph, the periodic fluctuation is not considered when calculating each difference value.

[0041] By adopting the technical solutions described above, the present application can bring the following beneficial effects:

[0042] Multi-modal fusion evaluation: the present application integrates multi-modal physiological signals such as breathing, electroencephalogram, heart rate, electromyogram, facial expression and eye movement to comprehensively reflect the physiological response caused by olfactory stimulation, overcome the limitations of single modal evaluation, and improve the evaluation accuracy and reliability; when the collection of single features of the present application is not ideal, the comprehensive evaluation method can also avoid the influence of the accuracy of the overall result;

[0043] Objective and quantitative evaluation: the present application can reduce the dependence on the subjective report of the subject, realize the objective and quantitative evaluation of the olfactory function, and is suitable for special groups such as children and patients with cognitive impairment.

[0044] Synchronous acquisition and accurate analysis: the synchronous control module of the present application ensures the accurate time correspondence of multi-modal signals and olfactory stimulation, and realizes in-depth analysis of the olfactory function by combining advanced signal processing and fusion algorithms.

[0045] Clinical application value: the present application can be used for early diagnosis of olfactory disorders, evaluation of therapeutic effect, and auxiliary screening of neurodegenerative diseases (such as Alzheimer's disease and Parkinson's disease), and has important clinical application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and other drawings can be obtained by those skilled in the art without creative labor.

[0047] Figure 1 A modular schematic diagram of a multi-modal physiological signal fusion olfactory function evaluation device in the specific embodiment of the present application;

[0048] Figure 2 A deployment mode schematic diagram of a multi-modal physiological signal fusion olfactory function evaluation device in the specific embodiment of the present application;

[0049] Figure 3 A flowchart of a multi-modal physiological signal fusion olfactory function evaluation method in the specific embodiment of the present application;

[0050] Figure 4 A respiratory waveform graph of a subject with normal olfaction when stimulated by a non-pleasurable gas;

[0051] 1, an olfactory stimulation module; 2, a respiratory signal acquisition unit; 3, a heart rate acquisition unit; 4, an electromyography acquisition unit; 5, a camera. DETAILED DESCRIPTION

[0052] The embodiments of the present disclosure will be described in detail below with reference to the drawings.

[0053] The embodiments of the present disclosure will be described in detail below with reference to the drawings.

[0054] It should be noted that various aspects of the embodiments described below are within the scope of the appended claims. It should be apparent that the aspects described herein can be implemented in a wide variety of forms and that any specific structure and / or function described herein is merely illustrative. Based on the teachings provided herein one skilled in the art will appreciate the various ways in which the aspects described herein can be implemented. It should be noted that the aspects described herein can be implemented independently of each other or in any combination.

[0055] It is also necessary to note that the drawings provided in the following embodiments only illustrate the basic concept of the present disclosure in a schematic manner, and only show the components related to the present disclosure in the drawings, not drawn according to the number, shape and size of the components when actually implemented, and the type, number and ratio of each component when actually implemented can be a random change, and the component layout type can also be more complex.

[0056] In addition, in the following description, specific details are provided in order to facilitate a thorough understanding of the examples. However, one skilled in the art will understand that the aspects described can be practiced without these specific details.

[0057] In one embodiment of the present application, a multi-modal physiological signal fusion olfactory function evaluation device is provided, as shown in Figure 1 The device includes an olfactory stimulation module 1, a multi-modal physiological signal acquisition module, a synchronous control module and a data processing and analysis module. The synchronous control module and the data processing and analysis module can be integrated in the same terminal device. The terminal device is preferably a computer that can perform model training and model running, and is equipped with a driver that can control each module and receive data from each module.

[0058] The olfactory stimulation module 1 acts on the subject to provide an olfactory stimulus;

[0059] The synchronous control module achieves time synchronization between the olfactory stimulation module 1 and the multi-modal physiological signal acquisition module;

[0060] The data processing and analysis module performs fusion analysis on the time-synchronized multi-modal features to establish an olfactory function evaluation model. The evaluation model is used to realize quantitative evaluation of olfactory function;

[0061] As shown in Figure 2 The multi-modal physiological signal acquisition module includes:

[0062] The respiratory signal acquisition unit 2 acts on the subject to collect the subject's respiratory rate, respiratory flow rate and respiratory pattern information;

[0063] The heart rate acquisition unit 3 uses electrocardiogram or PPG (Photoplethysmography) to act on the subject to collect the subject's heart rate variability;

[0064] The electromyography acquisition unit 4 acts on the subject to collect the subject's trapezius muscle electromyography signal. The electrode patch can be attached to the outside of the endpoint two centimeters between the seventh cervical vertebra and the shoulder peak.

[0065] The respiratory signal acquisition unit 2 of the present embodiment uses a small air flow sensor, which is attached near the nostrils by medical tape, Figure 2 which is partially blocked by the nasal catheter.

[0066] The clock module is arranged on each unit of the olfactory stimulation module 1 and the multi-modal physiological signal acquisition module in the embodiment, or the clock module is used for time service, and the acquired multi-modal physiological signals have time characteristics.

[0067] The breathing pattern and the brain electrical activity of the patient with olfactory dysfunction have characteristic changes: the peak frequency of inhalation of the patient with congenital anosmia is lower than that of a normal person in a wakeful state, and the coefficient of variation (CoV) of the breathing cycle is increased. The difference can be used to achieve an olfactory dysfunction classification accuracy of 83%, and the change is significantly positively correlated with the increase of the delta wave power of the brain electrical activity. Meanwhile, the airflow rate significantly affects the signal-to-noise ratio (SNR) of the olfactory event-related potential, the high airflow rate can improve the SNR and enhance the oscillation of the theta band, which indicates that the airflow regulation abnormality may be related to the olfactory dysfunction. In addition, compared with normal individuals, the patient with olfactory dysfunction has significant differences in the facial micro-expression, the electromyographic activity, the heart rate, the eye movement and the like when facing different odor molecule stimulations. The embodiment realizes the spatio-temporal synchronous acquisition and intelligent fusion of the multi-modal signals, and it is difficult to achieve comprehensive and accurate evaluation of the olfactory function.

[0068] In the embodiment, the olfactory stimulation module 1 includes an odorant bin, an airflow control system and a nasal catheter.

[0069] The odorant bin stores odorants, the odorants include water and two pleasant gases and one unpleasant gas that only activate the olfactory nerve, the odorants in the embodiment do not activate the trigeminal nerve, the pleasant gases are preferably vanillin and phenethyl alcohol, and the unpleasant gas is hydrogen sulfide.

[0070] The airflow control system includes an air compressor and an airflow converter, the air compressor is connected with the airflow converter through an air pipe, the airflow converter is connected with the odorant bin through an air pipe, the airflow converter is provided with a first air outlet and a second air outlet, the first air outlet outputs air, the second air outlet outputs air containing odorants, and the first air outlet and the second air outlet are connected with the nasal catheter through an air pipe.

[0071] The odorant bin in the embodiment adopts a transparent bottle body, which facilitates observation of the state of the odorants; the airflow converter includes a tee pipe, a first two-way electromagnetic valve and a second two-way electromagnetic valve, the electromagnetic valves are switched by a foot-operated conversion switch to realize the alternate output of air and air containing odorants. In the multi-modal physiological signal acquisition module, the breathing signal acquisition unit 2 adopts a high-precision flow sensor (such as an SDP3x pressure sensor), the electroencephalogram signal acquisition unit adopts a 10-20 system electrode arrangement, the heart rate acquisition unit 3 adopts a photoelectric heart rate sensor, the electromyographic acquisition unit 4 adopts a surface electromyographic electrode arrangement on the orbicularis oculi muscle and the orbicularis oris muscle, and the facial expression and eye movement acquisition unit adopts a high-definition camera 5 (frame rate ≥ 30 fps).

[0072] In the model training stage, the nasal catheter of the embodiment can be set as multiple groups, and airflow stimulation can be performed on multiple subjects at the same time, wherein the breathing signal acquisition unit 2 is arranged on the nasal catheter and includes a flow sensor and a humidity sensor; the odorant is a chemical substance or mixture with no toxicity to the body or a specific smell; at the same time, as a comparison, water vapor without color and smell can be directly used as a no-odorant comparison.

[0073] In the embodiment, the multi-modal physiological signal acquisition module further includes a facial expression and eye movement acquisition unit; the facial expression and eye movement acquisition unit includes a camera 5; the camera 5 is arranged on the front of the nasal catheter or an independent support, and is used to acquire facial expression changes and eye movement trajectories of the subject.

[0074] The facial expression and eye movement acquisition unit can perform expression and eye movement recognition based on an existing AI module that has been trained.

[0075] In the embodiment, the synchronization control module includes a clock synchronization unit, which synchronizes the clocks of the olfactory stimulation module 1 and each physiological signal acquisition unit in the multi-modal physiological signal acquisition module through an RS485 bus protocol, sets a synchronization time as t0, and starts from the t0 time to guide the period of the respiratory frequency detection period Th to coordinate the work of the olfactory stimulation module 1 and each physiological signal acquisition unit. The synchronization control module of the embodiment synchronizes the clocks of the olfactory stimulation module 1 and each physiological signal based on the clock information of the olfactory stimulation module 1 or the multi-modal physiological signal acquisition module, or synchronizes the clocks of the olfactory stimulation module 1 and each physiological signal after the synchronization control module itself provides time to the olfactory stimulation module 1 or the multi-modal physiological signal acquisition module.

[0076] In the embodiment, the data processing and analysis module includes a signal preprocessing unit, a feature extraction unit, and a fusion evaluation unit.

[0077] The signal preprocessing unit filters and denoises the collected multi-modal physiological signals, thereby eliminating interference and improving the effectiveness of information.

[0078] The feature extraction unit extracts respiratory features, electroencephalogram features, heart rate variability, electromyogram features, facial expression features, and eye movement features from the preprocessed signals, and draws multiple time sequence waveform graphs according to subjects, odorants, and feature types; at the same time, the feature extraction unit can perform some mathematical operations and difference operations to support the training sample requirements of the subsequent fusion evaluation unit.

[0079] The fusion evaluation unit adopts a machine learning algorithm or a deep learning model to perform fusion analysis on the multi-modal features collected by the feature extraction unit, and establishes the olfactory function evaluation model.

[0080] In this embodiment, the training samples of the fusion evaluation unit are obtained based on a quantitative method, and the quantification includes whether the subject has olfactory dysfunction, the type of olfactory stimulant for stimulating the subject, and the type of physiological characteristics.

[0081] In this embodiment, the respiratory characteristics include respiratory rate, tidal volume, and peak inspiratory flow rate.

[0082] The heart rate variability includes heart rate and LF / HF ratio.

[0083] The electromyographic characteristics include electromyographic amplitude and electromyographic frequency characteristics.

[0084] The facial expression characteristics include micro-changes in the zygomatic major muscle and the corrugator supercilii muscle; micro-changes in the movement of the orbicularis oris muscle and the orbicularis oculi muscle can be considered as auxiliary.

[0085] The eye movement characteristics include pupil diameter change and gaze time.

[0086] If the AI module is not used, the pupil diameter change can be accurately collected by separately configuring an infrared pupil instrument, and the gaze time can be accurately collected by separately configuring an eye tracking system.

[0087] In some embodiments, the present embodiment arranges an electroencephalogram signal collection unit, which includes a plurality of electrodes arranged at specific parts of the subject's scalp (such as Fz, Cz, Pz, C3, C4, etc.) for collecting olfactory stimulation-induced electroencephalogram signals (OERP).

[0088] The heart rate collection unit 3 of the present embodiment can be attached to the subject's chest or wrist.

[0089] The present embodiment synchronously collects multi-modal physiological signals such as respiratory rate, respiratory pattern, heart rate, electromyography, electroencephalography, facial expression, and eye movement, and combines olfactory stimulation to achieve comprehensive and objective evaluation of olfactory function.

[0090] Since the fusion evaluation unit of the present embodiment is a machine learning algorithm or a deep learning model, each physiological characteristic needs to be weighted in the training samples, thereby improving the training efficiency. The olfactory function evaluation model then performs quantitative evaluation on the olfactory function; the results of the quantitative evaluation include normal olfaction, reduced olfaction, and loss of olfaction.

[0091] In one embodiment, the present application also proposes a multi-modal physiological signal fusion olfactory function evaluation method implemented by the multi-modal physiological signal fusion olfactory function evaluation device of the above-mentioned embodiments, which comprises the following steps:Figure 3 As shown, it includes the following steps:

[0092] S101: Collect respiratory multimodal characteristic information from multiple samples of individuals with normal sense of smell and multiple samples of patients with loss of smell, respectively, with and without odorant.

[0093] S102: Analyze the multimodal physiological signal data of each sample, perform Fourier transform on the data to obtain time series waveforms of each modality under the conditions of presence and absence of odorant; obtain the frequency, amplitude and peak value of different modalities based on the waveforms;

[0094] S103: For each frequency characteristic, amplitude characteristic, and peak characteristic, in:

[0095] The difference and sum in the same waveform diagram

[0096] The difference between two waveforms with the same modal characteristics but different olfactory states

[0097] Data annotation is performed to obtain training samples;

[0098] S104: Train the olfactory function assessment model based on the training samples;

[0099] S105: Evaluate the non-quantitative subjects based on the trained olfactory function assessment model, and determine the olfactory state of the non-quantitative subjects.

[0100] In this embodiment, labeled data is used for model training. S101 and S102 collect multimodal sample data of normal and abnormal sense of smell. A waveform is generated for each modality of each sample. Since the physiological modalities are different, the waveforms are very different. The difference obtained by comparing the waveforms of different modalities cannot be used as effective data for labeling. In addition, the waveform of subjects with normal sense of smell will inevitably fluctuate after being stimulated by olfactory agents.

[0101] Therefore, this embodiment is based on two types of differences, which can be understood as: labeling the difference of specific data in the same waveform graph, and labeling the difference between two waveform graphs of the same olfactory agent for two subjects in different olfactory states. This avoids the problem of difficulty in comparing waveform graphs with different modal characteristics, and at the same time achieves the fusion judgment of multimodal features.

[0102] In this embodiment, when plotting the time-series waveform of respiratory characteristics:

[0103] If the deodorant starts at the peak of the waveform, this periodic fluctuation is not considered when calculating the differences.

[0104] like Figure 4The respiratory data waveform graphs of a group of normal olfactory subjects under multi-stage stimulation of non-pleasant gas are shown; wherein the small box is framed at the peak state just when the gas is released, and thus can be removed as data bad points during training.

[0105] The above merely provides the specific implementation of the present disclosure, but the protection scope of the present disclosure is not limited thereto, any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present disclosure, which shall be covered 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

1. A multi-modal physiological signal fusion-based olfactory function evaluation device, characterized in that, The device comprises an olfactory stimulation module, a multi-modal physiological signal acquisition module, a synchronous control module and a data processing and analysis module. The olfactory stimulation module acts on the subject to provide olfactory stimulus. The synchronous control module realizes time synchronization of the olfactory stimulation module and the multi-modal physiological signal acquisition module. The data processing and analysis module performs fusion analysis on the multi-modal features after time synchronization to establish an olfactory function evaluation model; the evaluation model is used to realize quantitative evaluation of olfactory function. The multi-modal physiological signal acquisition module comprises: A respiratory signal acquisition unit acting on the subject to collect the subject's respiratory rate, respiratory flow and respiratory pattern information; A heart rate acquisition unit acting on the subject to collect the subject's heart rate variability; An electromyography acquisition unit acting on the subject to collect the subject's trapezius muscle electromyography signal.

2. The multi-modal physiological signals fusion based olfactory function assessment apparatus according to claim 1, characterized in that, The olfactory stimulation module comprises an odorant bin, an airflow control system and a nasal catheter. The odorant bin stores olfactory stimulus; the olfactory stimulus comprises water and two pleasant gases and one unpleasant gas that only activate the olfactory nerve. The airflow control system comprises an air compressor and an airflow converter; the air compressor is connected to the airflow converter through an air pipe; the airflow converter is connected to the odorant bin through an air pipe; the airflow converter is provided with a first air outlet and a second air outlet; the first air outlet outputs air; the second air outlet outputs air containing olfactory stimulus; the first air outlet and the second air outlet are connected to the nasal catheter through an air pipe.

3. The multi-modal physiological signals fusion based olfactory function assessment apparatus according to claim 1, wherein, The multi-modal physiological signal acquisition module further comprises a facial expression and eye movement acquisition unit; the facial expression and eye movement acquisition unit comprises a camera; the camera is arranged on a separate support and is used to collect the subject's facial expression changes and eye movement trajectories.

4. The multi-modal physiological signals fusion based olfactory function assessment apparatus according to claim 1, wherein, The synchronous control module comprises a clock synchronization unit; the clock synchronization unit synchronizes the clocks of the olfactory stimulation module and each physiological signal acquisition unit in the multi-modal physiological signal acquisition module through an RS485 bus protocol; the synchronization time is set as t0; starting from t0, the respiratory rate detection period Th is used as the guide period to coordinate the work of the olfactory stimulation module and each physiological signal acquisition unit.

5. The multi-modal physiological signals fusion based olfactory function assessment apparatus according to claim 1, wherein, The data processing and analysis module comprises a signal preprocessing unit, a feature extraction unit and a fusion evaluation unit. The signal preprocessing unit performs filtering and denoising processing on the collected multi-modal physiological signals. The feature extraction unit extracts respiratory features, heart rate variability, electromyography features, facial expression features and eye movement features from the preprocessed signals. The fusion evaluation unit adopts a machine learning algorithm or a deep learning model to perform fusion analysis on the multi-modal features collected by the feature extraction unit to establish the olfactory function evaluation model.

6. The multi-modal physiological signal fusion olfactory function evaluation device according to claim 5, wherein The respiratory features comprise respiratory rate, tidal volume and peak inspiratory flow rate; The heart rate variability comprises heart rate LF / HF ratio; The electromyography features comprise electromyography amplitude and electromyography frequency characteristics; The facial expression features comprise the movement characteristics of the zygomatic major muscle and the corrugator supercilii muscle; The eye movement features comprise pupil diameter changes and gaze time.

7. The multi-modal physiological signals fusion based olfactory function assessment apparatus according to claim 1, wherein, The olfactory function evaluation model is used to realize quantitative evaluation on olfactory function; and the result of the quantitative evaluation includes normal olfaction, hyposmia and anosmia.

8. A method of multi-modal physiological signal fusion based olfactory function assessment, implemented by the multi-modal physiological signal fusion based olfactory function assessment device according to any one of claims 1-7, characterized in that, The method comprises the following steps: S101: Collecting breathing multi-modal feature information of samples of multiple normal olfaction persons and multiple anosmia patients respectively in the presence and absence of an olfactory agent; S102: Analyzing the multi-modal physiological signal data of each sample, obtaining time sequence waveform graphs of each modal feature in the presence and absence of the olfactory agent after performing Fourier transform on the data, and obtaining the frequency, amplitude and peak value of different modal features according to the waveform graphs; S103: Performing data labeling on the difference between each frequency feature, amplitude feature and peak value feature and the waveform graph and the difference between two waveform graphs of the same modal feature and different olfactory states to obtain training samples; S104: Training the olfactory function evaluation model based on the training samples; S105: Evaluating non-quantitative subjects based on the trained olfactory function evaluation model to judge the olfactory state of the non-quantitative subjects. In the S102, when drawing the time sequence curve waveform graph of the breathing feature: If the start time of releasing the olfactory agent is at the amplitude peak value of the curve waveform graph, the period fluctuation is not considered when calculating each difference. ​ 9. The multi-modal physiological signals fusion based olfactory function assessment method according to claim 8, characterized in that, ​ ​

Images