An electroencephalography paradigm for assessing speech levels in children with cochlear implants

Abstract

The application discloses an electroencephalogram (EEG) mode for evaluating speech level of children with cochlear implant, comprising the following steps: 1) selection of stimulation material and construction of EEG mode; 2) playing of EEG mode and collection of EEG signal; and 3) processing of EEG signal. On the basis of guaranteeing that standard stimulation can form effective short-time memory in the brain, the type of bias stimulation is increased, the type of speech that can be evaluated in the experiment is increased in the same time, and the test efficiency for evaluating speech level of children with cochlear implant is higher. In order to obtain effective MMN data in the test, the resistance of the EEG cap electrode is reduced by smearing conductive paste between the electrode and the scalp in the signal collection process, noise is reduced, and a suitable sampling frequency is set. In the signal processing process, the data needs to be segmented, on-line filtered, baseline corrected and superimposed, and the interference data is excluded, so that more reliable MMN data can be obtained in the test.

Description

Technical Field

[0001] This invention relates to the field of brain-computer interfaces, specifically an EEG paradigm for assessing the speech level of children with cochlear implants. Background Technology

[0002] Cochlear implants are the only means of helping deaf children regain auditory and speech functions. They are also a successful treatment for severe to profound sensorineural hearing loss and are currently the most successful and widely used neural prosthesis worldwide. Unlike hearing aids that amplify sound, cochlear implants are neural repair devices that directly stimulate the auditory nerve with electrical impulses to provide auditory input to the brain. This can help hearing-impaired patients, especially those with congenital deafness, regain partial hearing and promote their speech perception, vocalization, and language development. Because speech development after implantation is complex, timely and effective postoperative assessment of the child's speech level is crucial. Currently, methods used to assess the auditory and speech abilities of children who have received cochlear implants include speech audiometry and scales, brain imaging techniques, and electroencephalography (EEG).

[0003] Electroencephalography (EEG) is currently the most suitable method for objectively assessing the speech level of children with cochlear implants. This technology records the neurophysiological signals generated by brain activity through electrodes attached to the scalp, reflecting cognitive function. When the brain distinguishes different sounds, it generates a component of brainwaves called mismatch negativity (MMN). MMN can reflect the ability of children with cochlear implants to differentiate between different sounds; however, the generation of MMN requires a highly specific stimulus presentation paradigm, currently relying on the classic Oddball paradigm. In the classic Oddball paradigm, there are two stimuli (S1 and S2). S1 appears with a high probability of repetition (approximately 85-90%), forming short-term memory in the brain and is called the standard stimulus. Occasionally, a different stimulus, S2, appears; S2 is called the deviant stimulus. When the brain detects the occasional deviant stimulus within the standard stimulus (S1), S2 will induce an MMN response. The MMN response can be obtained by subtracting the S1-induced brainwave from the S2-induced brainwave. When the brain has difficulty distinguishing between different stimuli, the MMN response is absent or abnormal. However, the classic Oddball paradigm has a drawback: its efficiency is extremely low. The generation of the MMN (Mild Voice Noise) relies on biased stimuli, but these stimuli need to be repeated 100-200 times with a low probability to ensure a good signal-to-noise ratio in the EEG waveform and reliable results. Extrapolating from this, a single classic Oddball test would take at least ten minutes, and each experiment can only compare the ability to distinguish between two sounds. Therefore, an objective and efficient method is needed to assess the speech level of children with cochlear implants. Summary of the Invention

[0004] The purpose of this invention is to provide an EEG paradigm for assessing the speech level of children who have undergone cochlear implantation, so as to safely, objectively and efficiently assess the speech level of deaf children after cochlear implantation surgery, thereby solving the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] An EEG paradigm for assessing speech levels in children with cochlear implants includes the following steps:

[0007] 1) Selection of stimulation materials and construction of EEG paradigm

[0008] The stimulation materials consist of two types of sound stimuli: standard stimuli and deviated stimuli. The standard stimuli are randomly interspersed with at least two deviated stimuli with a moderate probability. The standard stimuli and each deviated stimuli appear randomly in each stimulus sequence at a ratio of at least N:1, where N is the number of deviated stimuli. An interleaved EEG paradigm is constructed.

[0009] 2) EEG paradigm playback and EEG signal acquisition

[0010] The software used to play the interleaved EEG paradigm program is used to present sound stimuli with a sound intensity range of 50-80dB. The EEG cap worn on the head for recording EEG signals is connected to an amplifier for amplifying and transmitting EEG signals, while the resistance of the EEG cap electrodes is reduced. EEG signals are synchronously acquired at a sampling frequency of 500-2000Hz.

[0011] 3) Processing of EEG signals

[0012] ① Connect the three sets of EEG data obtained from the test into a continuous EEG data segment and import it into the analysis software and segment the data; take the time when each sound stimulus appears as the starting point of the segment, and the number of segments is consistent with the number of times the stimulus is played;

[0013] ② Observe and remove bad segments that are obviously affected by electromyography and use relevant algorithms to remove cochlear implant artifacts;

[0014] ③ The signals induced by each stimulus are superimposed and averaged separately. The MMN waveform is obtained by subtracting the EEG waveform induced by the superimposed standard stimulus from the EEG waveform induced by the superimposed bias stimulus.

[0015] As a further aspect of the present invention, in step 1), the stimulus material consists of seven Chinese monosyllabic stimuli: / ta1 / , / tu1 / , / te1 / , / da1 / , / ra1 / , / ta2 / , and / ta4 / (the numbers represent the Chinese tones of the syllables). / ta1 / is the standard stimulus, while / tu1 / , / te1 / , / da1 / , / ra1 / , / ta2 / , and / ta4 / are the deviation stimuli. The stimulus material comprehensively includes vowels, consonants, and tones.

[0016] As a further aspect of the present invention, each sound stimulus lasts for 350ms, with a 650ms interval (silence period) after the sound ends, and then the next stimulus appears. One set of experiments contains 720 stimuli, including 360 standard stimuli and 60 deviated stimuli. Subjects need to conduct three sets of experiments, with each set of tests lasting 12 minutes and a 5-minute rest between sets.

[0017] As a further aspect of the present invention, the stimulation material is collected in an electromagnetically shielded soundproof room, and the beginning and end 5ms of the audio waveform are subjected to rising and falling envelope processing, respectively, to ensure consistent sound intensity.

[0018] As a further aspect of the present invention, in step 2), a conductive paste to reduce electrode resistance is injected between the electrodes of the EEG cap and the scalp.

[0019] As a further aspect of the present invention, in step 3), the data is filtered online at 0.05-100Hz before segmentation. The segment length of each sound stimulus segment is 400ms, including 100ms before the sound starts and 300ms after the sound starts. Baseline correction is performed after segmentation.

[0020] As a further aspect of the present invention, in step 3), the relevant algorithm is an independent component analysis algorithm.

[0021] As a further embodiment of the present invention, the EEG cap is also connected to a grounding electrode placed on the forehead, two left and right mastoid recording electrodes, two vertical electrooculography recording electrodes placed above and below the left eye, two horizontal electrooculography recording electrodes placed on the outer sides of both eyelids, and a reference electrode placed on the tip of the nose.

[0022] Compared with existing technologies, the beneficial effects of this invention are: while ensuring that standard stimuli can form effective short-term memories in the brain, it increases the types of biased stimuli, thereby increasing the types of speech that can be assessed in the experiment within the same time, making the test more efficient for assessing the speech level of children with cochlear implants; in order to obtain effective MMN data, during signal acquisition, it is necessary to reduce noise by applying conductive gel between the electrodes and the scalp and setting an appropriate sampling frequency; and during signal processing, it is necessary to perform steps such as data segmentation, online filtering, baseline correction, and superposition, and eliminate interference data, so that the test can obtain more reliable MMN data. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the interleaved Oddball EEG paradigm stimulation sequence;

[0024] Figure 2 MMN waveforms induced by the interleaved Oddball EEG paradigm in children with normal hearing and children with cochlear implants. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please see Figure 1 In this embodiment of the invention, an EEG paradigm for assessing the speech level of children with cochlear implants includes the following steps:

[0027] 1) Selection of stimulation materials and construction of EEG paradigm

[0028] The stimulation materials consist of two types of sound stimuli: standard stimuli and deviated stimuli. Standard stimuli are randomly interspersed with at least two deviated stimuli at a moderate probability in each sequence. The ratio of standard stimuli to each deviated stimulus is at least N:1 in each stimulus sequence, where N is the number of deviated stimuli. This constructs an interleaved EEG paradigm. A schematic diagram of the stimulation sequence is shown below. Figure 1 As shown, even though standard stimuli can form effective short-term memories in the brain, they can still constitute the Oddball paradigm and induce the MMN response, and the efficiency is N times that of the classic Oddball paradigm.

[0029] 2) EEG paradigm playback and EEG signal acquisition

[0030] Software was run to play an interleaved EEG paradigm program to present sound stimuli with a sound intensity range of 50-80dB. An EEG cap worn on the head to record EEG signals was connected to an amplifier to amplify and transmit the EEG signals, while the resistance of the EEG cap electrodes was reduced. EEG signals were synchronously acquired at a sampling frequency of 500-2000Hz.

[0031] 3) Processing of EEG signals

[0032] ① Connect the three sets of EEG data obtained from the test into a continuous EEG data segment and import it into the analysis software and segment the data; take the time when each sound stimulus appears as the starting point of the segment, and the number of segments is consistent with the number of times the stimulus is played;

[0033] ② Observe and remove bad segments that are obviously affected by electromyography and use relevant algorithms to remove cochlear implant artifacts;

[0034] ③ The signals induced by each stimulus are superimposed and averaged separately. The MMN waveform is obtained by subtracting the EEG waveform induced by the superimposed standard stimulus from the EEG waveform induced by the superimposed bias stimulus.

[0035] Furthermore, in step 1), the stimulus material consists of seven Chinese monosyllabic stimuli: / ta1 / , / tu1 / , / te1 / , / da1 / , / ra1 / , / ta2 / , and / ta4 / (the numbers represent the Chinese tones of the syllables). / ta1 / is the standard stimulus, while / tu1 / , / te1 / , / da1 / , / ra1 / , / ta2 / , and / ta4 / are the deviation stimuli. The stimulus material comprehensively includes vowels, consonants, and tones. China has one of the largest numbers of cochlear implant recipients in the world. However, because Chinese is a tonal language, the ability to distinguish different tones is crucial for Chinese children's vocabulary comprehension and the development of normal communication skills. Furthermore, different tonal combinations of the same consonants and vowels represent different meanings. Therefore, the assessment method and materials should differ from internationally accepted English-based scales. This testing method selected six deviation stimuli to comprehensively assess the three components of a Chinese monosyllabic syllable: vowels, consonants, and tones. Furthermore, among the six bias stimuli, three are syllables that do not exist in Chinese, meaning they have no corresponding Chinese characters or semantic pseudo-syllables: / te1 / , / ra1 / , and / ta2 / . The significance of this design lies in its ability to differentiate between the sensitivity to acoustic features and the impact of speech training. For example, children with cochlear implants may have difficulty recognizing tones, which could be due to the cochlear implant device not effectively improving their sound recognition, or it could be due to insufficient speech training, preventing them from processing their native language with the same skill as normal children. In this case, analyzing the MMN response evoked by the pseudo-syllables reveals that if the response is strong, it indicates that the device can process the acoustic features, and assistance should be sought from a rehabilitation institution to strengthen or improve speech training methods. Conversely, if the MMN response is weak, it indicates that the device's function is not fully utilized, and professional medical advice and consultation with the manufacturer are necessary.

[0036] Furthermore, each sound stimulus lasts for 350ms, followed by a 650ms interval (silence period) after the sound ends, and then the next stimulus is introduced. One set of experiments contains 720 stimuli, including 360 standard stimuli and 60 deviated stimuli. Subjects need to perform three sets of experiments, with each set lasting 12 minutes and a 5-minute rest between sets.

[0037] Furthermore, the stimulation material was collected in an electromagnetically shielded soundproof room, and the beginning and end 5ms of the audio waveform were processed with rising and falling envelopes, respectively, to ensure consistent sound intensity.

[0038] Furthermore, in step 2), a conductive paste to reduce electrode resistance is injected between the electrodes of the EEG cap and the scalp. The conductive paste effectively reduces noise during the testing process and ensures the clarity of the test results.

[0039] Furthermore, in step 3), the data is filtered online from 0.05 to 100 Hz before segmentation. Each segment of the segmented data, representing a sound stimulus, is 400 ms long, including 100 ms before the sound begins and 300 ms after the sound begins. Baseline correction is performed after segmentation. The purpose of filtering is to reduce noise during signal acquisition, and baseline correction can reduce deviations caused by data drift.

[0040] Furthermore, the relevant algorithm is an independent component analysis (ICA) algorithm. ICA is an effective method for removing cochlear implant artifacts. Principal component analysis is performed on the data, and the waveform, 2D topographic map, and power spectrum of each component are observed. The main characteristics of cochlear implant artifact components are: the rising and falling points of the waveform coincide with the start and end points of the sound; energy is concentrated at the location of the cochlear implant in the topographic map; the power spectrum shows high energy, 10-100 times that of normal EEG components. Cochlear implant artifacts are identified and removed from the EEG signal. In addition, segments with amplitudes exceeding ±100 μV should also be deleted.

[0041] Furthermore, the EEG cap is also connected to one ground electrode placed on the forehead, two left and right mastoid recording electrodes, two vertical electrooculography (EOG) recording electrodes placed above and below the left eye, two horizontal EOG recording electrodes placed on the outer sides of both eyelids, and one reference electrode placed on the tip of the nose. The ground electrode is used to eliminate interference. The mastoid electrodes and the nose tip electrode are both reference electrodes. All electrodes on the EEG cap are referenced to the reference electrodes. The horizontal and vertical EOG recording electrodes are used to monitor interference generated by eye movements such as blinking and eye movements.

[0042] The invention is used as follows:

[0043] 1. Preparation before testing: Wash head and face.

[0044] Before the test, the children should wash their hair with shampoo to remove oil and dandruff from the scalp and hair, thus effectively reducing scalp resistance. After washing, use a hairdryer to dry the hair and scalp to prevent signal drift caused by dampness. Then, apply a scrub to the tip of the nose, above the eyebrows, under the eyes, at the outer edge of both eyelids, and behind the ears using a cotton swab, gently wiping to remove oil and dandruff. Wipe clean with a tissue soaked in alcohol.

[0045] 2. Scalp electrode placement

[0046] After being cleaned, the subjects sat on a comfortable sofa. An EEG cap was then fitted: the cap, embedded with 64 silver / silver chloride electrodes, was worn on the subject's head to collect EEG signals from the scalp. The electrode placement followed the international 10-20 system. In addition, one ground electrode was placed on the forehead, two left and right mastoid recording electrodes, two vertical electrooculography (EOG) electrodes above and below the left eye, two horizontal EOG electrodes on the lateral aspects of both eyelids, and one reference electrode at the tip of the nose.

[0047] After putting on the EEG cap, insert the electronic parallel port at the end of the cap into the amplifier head box. The amplifier amplifies the recorded EEG signals and transmits them to the computer for storage. Open the EEG recording software on the computer and observe the resistance of each electrode. Inject conductive ointment (commercial product, mainly composed of starch and sodium chloride, with no health or safety risks) between the electrodes and the scalp to reduce resistance; the electrode resistance should be kept below 5kΩ.

[0048] 3. Cortical EEG signal recording

[0049] The E-prime software was run, playing an interleaved EEG paradigm program with sound stimuli at an intensity of 70 dB SPL. Cortical EEG signals were simultaneously acquired. The sampling rate was set to 500 Hz, with online filtering from 0.05 to 100 Hz. During the test, a screen was placed in front of the child, playing a silent animated film. The child only needed to watch the silent film; there were no test-related tasks, and the child was not required to focus on the stimulus itself. Each test lasted 12 minutes, with a total of 3 tests conducted, followed by a 5-minute rest period between each set.

[0050] 4. Test ends

[0051] After the test, wash your hair again with shampoo and blow-dry it. The whole process is now complete.

[0052] 5. Test Result Analysis

[0053] The test results were evaluated based on the MMN waveform. Cochlear implant recipients with different speech recognition abilities exhibited different MMN morphologies. The amplitude and latency of the MMN waveform in normal control children were used as the evaluation criteria. This test contained six types of bias, resulting in six different MMN waveforms. MMN amplitude was calculated by taking a 20ms window before and after the peak of the MMN response, and the average amplitude of the MMN within this 40ms window was taken as the MMN amplitude. The time point corresponding to the peak of the MMN response was the MMN latency. Cochlear implant recipients with good speech recognition abilities showed larger MMN amplitudes and shorter latencies, similar to the MMN amplitudes of normal control children. In contrast, cochlear implant recipients with moderate to low speech recognition abilities showed smaller MMN amplitudes and longer latencies, or even no MMN at all. Figure 2As shown, the first and second rows present the vowel results; the third and fourth rows present the consonant results; and the fifth and sixth rows present the tone results. The gray rectangles marking the time points corresponding to the MMN response peaks indicate the MMN responses. The Oddball paradigm can induce MMN responses in both normal and hearing-normal children, demonstrating the effectiveness of the method. Compared to hearing-normal children (control group), children with cochlear implants showed lower MMN response amplitudes and longer latency. These results suggest that speech discrimination ability in children with cochlear implants is inferior to that of their age-normal peers.

[0054] Although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0055] Therefore, the above description is only a preferred embodiment of this application and is not intended to limit the scope of this application; that is, all equivalent modifications made in accordance with the scope of the claims of this application shall be within the protection scope of the claims of this application.

Claims

1. An EEG paradigm for assessing the speech level of children with cochlear implants, characterized in that, Includes the following steps: 1) Selection of stimulation materials and construction of EEG paradigm The stimulation materials consist of two types of sound stimuli: standard stimuli and deviated stimuli. The standard stimuli are randomly interspersed with at least two deviated stimuli with a moderate probability. The standard stimuli and each deviated stimuli appear randomly in each stimulus sequence at a ratio of at least N:1, where N is the number of deviated stimuli. An interleaved EEG paradigm is constructed. In step 1), the stimulus material consists of seven Chinese monosyllabic stimuli: / ta1 / , / tu1 / , / te1 / , / da1 / , / ra1 / , / ta2 / , and / ta4 / , with the numbers representing the Chinese tones of the syllables. / ta1 / is the standard stimulus, while / tu1 / , / te1 / , / da1 / , / ra1 / , / ta2 / , and / ta4 / are the deviation stimuli. The stimulus material comprehensively includes vowels, consonants, and tones. Each sound stimulus lasts for 350ms, followed by a 650ms interval before the next stimulus appears. One set of experiments contains 720 stimuli, including 360 standard stimuli and 60 deviance stimuli. Subjects are required to complete three sets of experiments, with each set lasting 12 minutes. 2) EEG paradigm playback and EEG signal acquisition The software used to play the interleaved EEG paradigm program is used to present sound stimuli with a sound intensity range of 50-80dB. The EEG cap worn on the head for recording EEG signals is connected to an amplifier for amplifying and transmitting EEG signals, while the resistance of the EEG cap electrodes is reduced. EEG signals are synchronously acquired at a sampling frequency of 500-2000Hz. 3) Processing of EEG signals ① Connect the three sets of EEG data obtained from the test into a continuous EEG data segment and import it into the analysis software and segment the data; take the time when each sound stimulus appears as the starting point of the segment, and the number of segments is consistent with the number of times the stimulus is played; ② Observe and remove bad segments that are obviously affected by electromyography and use relevant algorithms to remove cochlear implant artifacts; ③ The signals induced by each stimulus are superimposed and averaged separately. The MMN waveform is obtained by subtracting the EEG waveform induced by the superimposed standard stimulus from the EEG waveform induced by the superimposed bias stimulus.

2. The EEG paradigm for assessing the speech level of children with cochlear implants according to claim 1, characterized in that, The stimulation material was collected in an electromagnetically shielded, soundproof room, and the audio waveform was subjected to rising and falling envelope processing at the beginning and end 5ms, respectively, with the sound intensity adjusted to be consistent.

3. The EEG paradigm for evaluating the speech level of children with cochlear implants according to claim 1, characterized in that, In step 2), a conductive paste to reduce electrode resistance is injected between the electrodes of the EEG cap and the scalp.

4. The EEG paradigm for evaluating the speech level of children with cochlear implants according to claim 1, characterized in that, In step 3), the data is filtered online from 0.05 to 100 Hz before segmentation. The segment length of each sound stimulus segment is 400 ms, including 100 ms before the sound starts and 300 ms after the sound starts. Baseline correction is performed after segmentation.

5. The EEG paradigm for evaluating the speech level of children with cochlear implants according to claim 1, characterized in that, In step 3), the relevant algorithm is an independent component analysis algorithm.

6. The EEG paradigm for assessing the speech level of children with cochlear implants according to claim 1, characterized in that, The EEG cap is also connected to a ground electrode placed on the forehead, two left and right mastoid recording electrodes, two vertical EEG recording electrodes placed above and below the left eye, two horizontal EEG recording electrodes placed on the outer sides of both eyelids, and a reference electrode placed on the tip of the nose.

Images