The processing mechanism of the human brain for speech information is a significant source of inspiration for the study of speech enhancement technology. Attention and lateral inhibition are key mechanisms in auditory information processing that can selectively enhance specific information. Building on this, the study introduces a dual-branch U-Net that integrates lateral inhibition and feedback-driven attention mechanisms. Noisy speech signals input into the first branch of the U-Net led to the selective feedback of time-frequency units with high confidence. The generated activation layer gradients, in conjunction with the lateral inhibition mechanism, were utilized to calculate attention maps. These maps were then concatenated to the second branch of the U-Net, directing the network's focus and achieving selective enhancement of auditory speech signals. The evaluation of the speech enhancement effect was conducted by utilising five metrics, including perceptual evaluation of speech quality. This method was compared horizontally with five other methods: Wiener, SEGAN, PHASEN, Demucs and GRN. The experimental results demonstrated that the proposed method improved speech signal enhancement capabilities in various noise scenarios by 18% to 21% compared to the baseline network across multiple performance metrics. This improvement was particularly notable in low signal-to-noise ratio conditions, where the proposed method exhibited a significant performance advantage over other methods. The speech enhancement technique based on lateral inhibition and feedback-driven attention mechanisms holds significant potential in auditory speech enhancement, making it suitable for clinical practices related to artificial cochleae and hearing aids.
Humans ; Attention/physiology* ; Speech Perception/physiology* ; Neural Networks, Computer ; Speech ; Noise ; Feedback

Musical training can counteract age-related decline in speech perception in noisy environments. However, it remains unclear whether older non-musicians and musicians rely on functional compensation or functional preservation to counteract the adverse effects of aging. This study utilized resting-state functional connectivity (FC) to investigate functional lateralization, a fundamental organization feature, in older musicians (OM), older non-musicians (ONM), and young non-musicians (YNM). Results showed that OM outperformed ONM and achieved comparable performance to YNM in speech-in-noise and speech-in-speech tasks. ONM exhibited reduced lateralization than YNM in lateralization index (LI) of intrahemispheric FC (LI_intra) in the cingulo-opercular network (CON) and LI of interhemispheric heterotopic FC (LI_he) in the language network (LAN). Conversely, OM showed higher neural alignment to YNM (i.e., a more similar lateralization pattern) compared to ONM in CON, LAN, frontoparietal network (FPN), dorsal attention network (DAN), and default mode network (DMN), indicating preservation of youth-like lateralization patterns due to musical experience. Furthermore, in ONM, stronger left-lateralized and lower alignment-to-young of LI_intra in the somatomotor network (SMN) and DAN and LI_he in DMN correlated with better speech performance, indicating a functional compensation mechanism. In contrast, stronger right-lateralized LI_intra in FPN and DAN and higher alignment-to-young of LI_he in LAN correlated with better performance in OM, suggesting a functional preservation mechanism. These findings highlight the differential roles of functional preservation and compensation of lateralization in speech perception in noise among elderly individuals with and without musical expertise, offering insights into successful aging theories from the lens of functional lateralization and speech perception.
Humans ; Speech Perception/physiology* ; Music ; Male ; Functional Laterality/physiology* ; Female ; Aged ; Adult ; Young Adult ; Aging/physiology* ; Middle Aged ; Magnetic Resonance Imaging ; Brain/physiology*

Objective: Functional near-infrared spectroscopy (fNIRS) was used to study the effect of aging on the neuroimaging characteristics of cerebral cortex in the process of speech perception. Method: Thirty-four adults with normal hearing were recruited from March 2021 to June 2021, including 17 in the young group, with 6 males, 11 females, age (32.1±5.0) years, age range 20-39 years. and 17 in the elderly group, with 6 males, 11 females, age (63.2±2.8) years, age range 60-70 years. The test material was the sentence table of the Mandarin Hearing Test in Noise (MHINT). The task state block experiment design was adopted, and the temporal lobe, Broca's area, Wernicke's area, motor cortex were used as regions of interest. Objective brain imaging technology (fNIRS) combined with subjective psychophysical testing method was used to analyze the activation area and degree of cerebral cortex related to auditory speech perception in the elderly and young people under different listening conditions (quiet, signal-to-noise ratio of 10 dB, 5 dB, 0 dB, -5 dB). SPSS 23 software was used for statistical analysis. Result: The activation area and degree of activation in the elderly group were lower than those in the young group under each task condition; The number of activation channels in the young group were significantly more than those in the old group, and the number of activation channels in the left hemisphere were more than those in the right hemisphere, but there was no difference in the number of activation channels. There were more channels affected by age in the left hemisphere than in the right hemisphere. The activation degree of the young group when the signal-to-noise ratio was 0 dB was significantly higher than that of other signal-to-noise ratio conditions (P<0.05), but there was no significant difference in the old group under the five conditions (P>0.05). The speech recognition score of the young group was higher than that of the old group under all conditions. When the quiet and signal-to-noise ratio was 10 dB, the correct score of the two groups was equal or close to 100%. With the gradual decrease of signal-to-noise ratio, there was a significant difference between the two groups when the signal-to-noise ratio was 5 dB. The speech recognition accuracy of the young group decreased significantly when the signal-to-noise ratio was 0 dB, while that of the old group decreased significantly when the signal-to-noise ratio was 5 dB. Conclusions: With the increase of age, the speech perception in noisy environment and the activity of cerebral cortex gradually deteriorate, and the speech dominance hemisphere (left hemisphere) will be significantly affected by aging. The overall activation area and activation degree of the elderly under different speech tasks are lower than those of the young.
Acoustic Stimulation/methods* ; Adolescent ; Adult ; Aged ; Auditory Cortex/physiology* ; Female ; Humans ; Male ; Middle Aged ; Spectroscopy, Near-Infrared ; Speech Perception/physiology* ; Technology ; Young Adult

Speech comprehension is a central cognitive function of the human brain. In cognitive neuroscience, a fundamental question is to understand how neural activity encodes the acoustic properties of a continuous speech stream and resolves multiple levels of linguistic structures at the same time. This paper reviews the recently developed research paradigms that employ electroencephalography (EEG) or magnetoencephalography (MEG) to capture neural tracking of acoustic features or linguistic structures of continuous speech. This review focuses on two questions in speech processing: (1) The encoding of continuously changing acoustic properties of speech; (2) The representation of hierarchical linguistic units, including syllables, words, phrases and sentences. Studies have found that the low-frequency cortical activity tracks the speech envelope. In addition, the cortical activities on different time scales track multiple levels of linguistic units and constitute a representation of hierarchically organized linguistic units. The article reviewed these studies, which provided new insights into the processes of continuous speech in the human brain.
Acoustic Stimulation ; Electroencephalography ; Humans ; Magnetoencephalography ; Speech ; physiology ; Speech Perception

Facial and vocal expressions are essential modalities mediating the perception of emotion and social communication. Nonetheless, currently little is known about how emotion perception and its neural substrates differ across facial expression and vocal prosody. To clarify this issue, functional MRI scans were acquired in Study 1, in which participants were asked to discriminate the valence of emotional expression (angry, happy or neutral) from facial, vocal, or bimodal stimuli. In Study 2, we used an affective priming task (unimodal materials as primers and bimodal materials as target) and participants were asked to rate the intensity, valence, and arousal of the targets. Study 1 showed higher accuracy and shorter response latencies in the facial than in the vocal modality for a happy expression. Whole-brain analysis showed enhanced activation during facial compared to vocal emotions in the inferior temporal-occipital regions. Region of interest analysis showed a higher percentage signal change for facial than for vocal anger in the superior temporal sulcus. Study 2 showed that facial relative to vocal priming of anger had a greater influence on perceived emotion for bimodal targets, irrespective of the target valence. These findings suggest that facial expression is associated with enhanced emotion perception compared to equivalent vocal prosodies.
Adult ; Brain Mapping ; methods ; Cerebral Cortex ; diagnostic imaging ; physiology ; Emotions ; physiology ; Facial Expression ; Facial Recognition ; physiology ; Female ; Humans ; Magnetic Resonance Imaging ; Psychomotor Performance ; physiology ; Social Perception ; Speech Perception ; physiology ; Young Adult

OBJECTIVE: To analyze the short monosyllabic list of Mandarin speech test materials (MSTMs) which have been evaluated the equivalence of difficulty, and to establish the performance-intensity function (P-I function) for people with normal hearing as clinical reference of hearing recovery and individuals ability to perceive and process speech. METHOD: Thirty-seven subjects (the age ranged from 18 to 26 years old) who speak Mandarin well in their daily lives with normal hearing participated in this study. Eight lists of the Short Mandarin Monosyllabic materials (20 words per list) with equal difficulty were utilized. The results were analyzed by Statistical Package for the Social Sciences (SPSS) software version 17.0. RESULT: P-I function for short monosyllabic word list was x = 98.557/(1 + 12.243 exp (-0.17(P-15, x(max) = 98.557. And the sound pressure level of speech corresponding to a 50% recognition score was 29.6 dB SPL or 9.6 dB HL. The results showed P-I function of 3.1 per dB for Mandarin materials. CONCLUSION The study established the P-I function of the Mandarin short monosyllabic word list materials with equal difficulty, which provides the normative data for identifying the normal hearing in a clinical setting.
Adult ; Auditory Perception ; Female ; Hearing Tests ; methods ; Humans ; Male ; Speech ; Speech Perception ; physiology

The auditory steady state response (ASSR) may reflect activity from different regions of the brain, depending on the modulation frequency used. In general, responses induced by low rates (≤40 Hz) emanate mostly from central structures of the brain, and responses from high rates (≥80 Hz) emanate mostly from the peripheral auditory nerve or brainstem structures. Besides, it was reported that the gamma band ASSR (30-90 Hz) played an important role in working memory, speech understanding and recognition. This paper investigated the 40 Hz ASSR evoked by modulated speech and reversed speech. The speech was Chinese phrase voice, and the noise-like reversed speech was obtained by temporally reversing the speech. Both auditory stimuli were modulated with a frequency of 40 Hz. Ten healthy subjects and 5 patients with hallucination symptom participated in the experiment. Results showed reduction in left auditory cortex response when healthy subjects listened to the reversed speech compared with the speech. In contrast, when the patients who experienced auditory hallucinations listened to the reversed speech, the auditory cortex of left hemispheric responded more actively. The ASSR results were consistent with the behavior results of patients. Therefore, the gamma band ASSR is expected to be helpful for rapid and objective diagnosis of hallucination in clinic.
Acoustic Stimulation ; methods ; Adult ; Algorithms ; Auditory Cortex ; physiopathology ; Auditory Perception ; physiology ; Electroencephalography ; methods ; Female ; Hallucinations ; diagnosis ; physiopathology ; Humans ; Linguistics ; Male ; Middle Aged ; Models, Neurological ; Speech ; Young Adult

Speech evoked brainstem responses (s-ABRs) elicited by a speech syllable /da/ are composed of four parts: onset response (OR), transitional response, frequency following response (FFR) and offset response. FFR elicited by periodic events behaves like a quasi-periodic waveform corresponding to the stimulus sounds. The fast Fourier transform based spectra are commonly used to exam the characteristics of s-ABR in practice, which is, however, unable to trace the occurrence of the main components of s-ABR. The FFR is usually not obvious in the original individual s-ABR waveform. In this paper, we proposed a novel approach to observe the FFR by an instantaneous energy spectrum performed on the intrinsic mode functions (IMFs) after empirical mode decomposition (EMD) of the s-ABR. We demonstrated that the FFR is most pronounced on the second layer of IMFs. This finding suggests a new way which may be available to characterize and to detect the FFR better. This will benefit the clinic applications of s-ABRs.
Adult ; Brain Stem ; physiology ; Evoked Potentials, Auditory, Brain Stem ; physiology ; Female ; Fourier Analysis ; Humans ; Male ; Speech ; Speech Perception ; physiology ; Young Adult

BACKGROUNDMany factors interfering with a listener attempting to grasp speech in noisy environments. The spatial hearing by which speech and noise can be spatially separated may play a crucial role in speech recognition in the presence of competing noise. This study aimed to assess whether, and to what degree, spatial hearing benefit speech recognition in young normal-hearing participants in both quiet and noisy environments.

METHODSTwenty-eight young participants were tested by Mandarin Hearing In Noise Test (MHINT) in quiet and noisy environments. The assessment method used was characterized by modifications of speech and noise configurations, as well as by changes of speech presentation mode. The benefit of spatial hearing was measured by speech recognition threshold (SRT) variation between speech condition 1 (SC1) and speech condition 2 (SC2).

RESULTSThere was no significant difference found in the SRT between SC1 and SC2 in quiet. SRT in SC1 was about 4.2 dB lower than that in SC2, both in speech-shaped and four-babble noise conditions. SRTs measured in both SC1 and SC2 were lower in the speech-shaped noise condition than in the four-babble noise condition.

CONCLUSIONSpatial hearing in young normal-hearing participants contribute to speech recognition in noisy environments, but provide no benefit to speech recognition in quiet environments, which may be due to the offset of auditory extrinsic redundancy against the lack of spatial hearing.

Adolescent ; Adult ; Auditory Threshold ; physiology ; Female ; Hearing ; physiology ; Humans ; Male ; Noise ; Speech Perception ; physiology ; Speech Reception Threshold Test ; Young Adult

To enhance speech recognition in noise, as well as tone recognition, we presented a new kind of speech coding strategy, called one-octave wavelet transform zero-crossing stimulation (WTZS), for cochlear implants based on amplitude and frequency modulation. We selected 15 volunteers with normal hearing ability to carry out hearing simulation experiments by picking up the amplitude (amplitude modulation, AM), zero-crossings (frequency modulation, FM) and gradient parameters from processed speech signal in the domain of one-octave wavelet transform to synthesize the stimulating pulstile series. The experimental results demonstrated that the phonetic recognition in quiet surroundings with amplitude modulation only strategy (CIS) is similar to that of amplitude and frequency modulations strategies (FAME and WTZS), while the tone perception of CIS is inferior to that of FAME and WTZS strategies. However, in noisy environment, the phonetic recognition, tone perception, as well as sentence recognition of WTZS strategy are better than those of CIS and FAME strategies. WTZS strategy, utilizing amplitude (AM), zero-crossings (FM) and gradient parameters to synthesize stimulus, can enhance the phonetic and tonal language recognition in noise environment effectively, and could be used in cochlear implant system for speech processor design after arithmetic optimization.
Cochlear Implants ; Humans ; Language ; Prosthesis Design ; Signal Processing, Computer-Assisted ; Sound Spectrography ; methods ; Speech Acoustics ; Speech Perception ; physiology