phone: (86-10) 6275-6804
09/1989 - 07/1994 Ph.D. (Department of Psychology, Carleton University)
09/1985 - 07/1988 M.Sc. (Department of Psychology, Peking University)
09/1981 - 07/1985 B.Sc. (Department of Psychology, Peking University)
EDITORIAL ADVISORY BOARD SERVICES
|08/2007 - Present||Member of Editorial Board of Neuroscience and Biobehavioral Reviews (the official journal of the International Behavioral Neuroscience Society)|
PROFESSIONAL JOURNAL REVIEW ACTIVITIES
- Acta Acustica United with Acustica；
- Brain Research；
- Brain Structure and Function；
- Canadian Acoustics；
- Canadian Journal of Neurological Sciences；
- Cerebral Cortex
- Clinical Neurophysiology；
- Clinical Psychopharmacology and Neuroscience；
- Ear and Hearing；
- European Journal of Neuroscienece；
- Hearing Research；
- International Journal of Developmental Neuroscience；
- Journal of Comparative Physiology；
- Journal of Neurophysiology；
- Neuroscience and Biobehavioral Reviews；
- Neuroscience Letters；
- Physiology & Behavior；
- PLoS ONE；
- Progress in Neuro-Psychopharmacology & Biological Psychiatry；
- Pharmacology，Biochemistry and Behavior；
- Psychological Medicine；
- Science in China Series C: Life Sciences；
- Progress in Biochemistry and Biophysics（in Chinese）；
- Acta Biophysica Sinica（in Chinese）；
- Advances in Psychological Science（in Chinese）；
- Acta Psychologica Sinica（in Chinese）；
- Science China；
- National Medical Journal of China
- Scientific Reports。
My research is focused on developing a fuller understanding of the interaction between auditory processing and auditory sensory gating. Of particular interest are neural substrates in the auditory brainstem and forebrain that contribute to sound localization, prepulse inhibition of the startle reflex, the precedence effect, and effects of auditory grouping on informational masking. The experiments are broadly based in neurophysiological, neuroanatomical, immunochemical, and psychophysical/behavioral techniques. For studies using animal subjects, electrophysiological works are often accomplished in whole animals and makes extensive use of single-unit and focal field potential recordings. Combinations of axonal tracing and immunochemical techniques are used to determine the functional connections of the neural circuits under studies. Computer-controlled automatic mazes are used to test visual and auditory discrimination performances. Current human projects involve the psychophysical studies on (1) the acoustic conditions that lead to fusion of sound gaps, (2) spatial release of acoustic informational (speech) masking, and (3) the role of acoustic flankers in attenuating the bounce-inducing effect. The advanced Tucker-Davis Technology System is used to create acoustic stimuli, run experimental procedures, and collect data. In addition to these projects on auditory processing and gating, my research extends to the limbic-cortical structures that contribute to startle modulation, emotional leaning, reversal learning, and motive-motor gating. I am also interested in animal models of schizophrenia. My research emphasizes the combination of the three complementary approaches: (1) establishing appropriate parallel human and animal psychophysical/behavioral models; (2) conducting comprehensive electrophysiological, neuroanatomical, immunochemical and even genetic experiments in animals, in research for multi-facet mechanisms; (3) building the theoretical models for conceptual explanation of the experimental discoveries.
TEN MOST SIGNIFICANT SCIENTIFIC CONTRIBUTIONS
(1) The temporal storage of acoustic information and the aging effect
Under a reverberant condition, listeners receive both the wavefront directly coming from a sound source and reflections of the source. If the delay time between the direct wave and a reflected wave (which is highly correlated with the source) is sufficiently short, signals of the direct sound wave can be perceptually integrated with those of the delayed reflection. For example, listeners can typically perceive a single “fused” sound image as coming from the site of the source without perceiving the distinct reflection coming from a location that is different from the source location. This phenomenon is called “the precedence effect”. However, because there is always a temporal gap between the source and the reflection, there must be certain mechanisms in the auditory system for bridging the temporal gap in achieving the lead-lag integration.
Using the transient drop in interaural correlation between two correlated broadband noises at the two ears as the probe signal, we found that fine-structure information of broadband noises can be retained for over 20 ms in some younger listeners. This long-lasting information persistence cannot be explained by either the delay-line model or the peripheral-filter model. Our explanation is that when the auditory system receives sounds, fine-structure information of the received sounds can be faithfully stored for a short period of time, and during this period of time correlations between sound waves are calculated. This pre-attentive signal buffer at the central level must also be critical for higher-order central processes. Moreover, results of our psychoacoustic and ERP experiments indicate that the temporal storage of acoustic details is frequency dependent: for narrowband noises, the temporal storage declines as the center frequency becomes higher. We also found that the temporal storage of acoustic details is significantly reduced in older-adult listeners with clinically normal hearing. We propose that the age-related reduction of the ability to temporally store acoustic fine structures is related to the age-related difficulties in identifying sound sources in noisy, reverberant environments.
(2) Attribute capture
Results of our psychoacoustic and ERP-recording experiments indicate that the precedence effect is based on perceptual capture of attributes of the reflection. When a lagging sound is highly correlated with a leading sound (like a reflection highly correlated with its source), non-spatial attributes of the lagging sound are perceptually captured, but not suppressed, by the leading sound. Different attributes of the lagging sound have different capture thresholds. With the decrease of the delay time between the correlated leading and lagging sounds, more and more attributes of the lagging sound are perceptually captured (the loudness of the lagging sound progressively becomes weaker). When all the attributes of the lagging sound are perceptually captured by the leading sound, listener will not perceive any sound image as coming from the physical location of the lagging sound source but a single fused sound image as coming from the location of the leading sound source. Thus perceptual fusion of the leading and lagging sounds is not an “all-or-none” phenomenon and based on higher-order central processing.
(3) The effect of perceived spatial separation on release speech from informational masking
Under cocktail party conditions, listeners usually find it difficult to comprehend and participate in conversations, especially when other people are talking. Two factors contribute to this difficulty: 1) energetic masking of the target speech by non-target sounds, and 2) informational interference from irrelevant talkers (informational masking). We found that in both younger adults and older adults, perceived spatial separation induced by the precedence effect, does not substantially change the acoustics at the ears, but can specifically release speech (English or Chinese) from informational masking. Thus the precedence effect is important for reducing informational masking in noisy, reverberant listening conditions by perceptually grouping correlated sounds and segregating uncorrelated sounds.
Interestingly, although the perceived spatial separation between target speech and masker causes similar releases from energetic masking for English speech and Chinese speech, it causes a larger release from informational masking for English speech than for Chinese speech, suggesting that Chinese, a tonal language with larger pitch-contour fluctuations, is more resistant to informational masking than English.
Our recent studies have also shown that there is a significant correlation between the ability to temporally retain low-frequency (£ 400 Hz) fine-structure information and the ability to release speech from informational masking through attribute capture under a simulated reverberant condition. Moreover, the ability to release speech from informational masking at long source/reflection delays for target speech is significantly lower in older-adult listeners than in younger-adult listeners. This age difference, we think, is partially due to the aging difference in temporal storage of acoustic details.
(4) Effects of non-spatial cuing on releasing speech from informational masking
In a cocktail-party environment, human listeners are able to use both perceptual-level and cognitive-level cues to segregate the attended target speech from other background conversations. We found that at the cognitive level, priming the listener with part of the target Chinese speech in quiet can markedly improve the recognition of the remaining parts when the target speech and competing speech are presented at the same time. In addition, familiarity or knowledge of the voice characteristics of the target talker can also help the listener attend to the target talker when other talkers are present. These results suggest that both content cues and voice cues can be used by listeners to release speech from informational masking.
Moreover, we found that listeners can also use speech-synchronized visual cues and speech-rate cues to significantly reduce informational masking of target speech. Thus we propose that in a cocktail-party environment, listeners can use various cues to facilitate perceptual segregation between target speech and masking speech, and how to integrate these cues under a specific situation should be under even-higher cognitive controls.
(5) Simulated phase-locking stimulation: an improved speech processing strategy for cochlear implants
Traditional speech-processing strategies used for cochlear implants, such as the Continuous Interleaved Sampling (CIS) strategy, extract only envelope information of narrow-band signals without utilizing phase information contained in sound signals. We have developed a new speech-processing strategy, the Simulated Phase-locking Stimulation (SPLS) strategy. In addition to extracting amplitude envelope from various frequency bands after band-pass filtering, this new strategy can extract phase information in speech. The advantages of the SPLS over the CIS are well confirmed by results of our psychoacoustic experiments, in which recognition of Chinese speech processed by the CIS strategy was compared with that of the SPLS strategy in normal-hearing listeners under either noise masking condition or speech masking condition. The new SPLS strategy is potentially useful in improving the speech-recognition performance of cochlear-implant users under cocktail party situations.
(6) Neural mechanisms underlying sensorimotor gating
Detection of a non-startling acoustic event that is presented shortly before an intense startling sound can lead to inhibition of the acoustic startle reflex in both humans and animals. This phenomenon is called “prepulse inhibition”, and widely used as a model of sensorimotor gating. The prepulse inhibition reflects a gating function of reducing influence of disruptive inputs at the early (pre-attentive) stages of sensory processing. In humans, prepulse inhibition can be modulated by both attention and emotional responses to the prepulse, indicating that the early-stage gating processing can be "top-down" regulated by higher-order cognitive processes.
Using laboratory animals (rats), we have thoroughly studied the neural pathways that mediate acoustic startle, the neural pathways that mediate prepulse inhibition, the neural pathways that modulate prepulse inhibition, and the neural pathways that mediate auditory-conditioning potentiation of startle. Based on results of our systematic experimental studies and previous studies by other investigators, we have published three important review articles to summarize the current knowledge on the neural mechanisms underlying prepulse inhibition. These three reviews have been published in Psychopharmacology (2001), Neuroscience and Biobehavioral Reviews (2002), andHearing Research (2002), respectively.
Our recent animal studies have for the first time shown that fear conditioning can modulate prepulse inhibition in rats. After a neutral acoustic prepulse is time-precisely combined with footshock, the prepulse becomes conditioned (biologically significant) and causes enhanced prepulse inhibition. Moreover, the conditioning-induced enhancement of prepulse inhibition depends on activity of metabotropic glutamate subtype 5 receptors. We have evidence to show that metabotropic glutamate subtype 5 receptors contribute to the early-phase of the long-term potentiation (LTP) recorded in the lateral nucleus of the amygdala under in vivo preparations. This new model for studying cognitive regulation of prepulse inhibition is also useful for studying both the cognitive and biological bases of schizophrenia (also see “(10) Animal models of schizophrenia”).
(7) Top-down modulation of rats’ prepulse inhibition
Using prepulse inhibition as the behavioral model for studying rats’ detection of transient acoustic signals, we found that rats are able to detect a sudden drop in inter-sound correlation with the considerably short duration threshold. These results suggest that rats can perceptually integrate correlated sounds waves, just like humans. In addition, the rat’s detection of a sudden drop in inter-sound correlation can be improve by both perceptual learning and emotional learning, indicating that auditory processing of inter-sound correlations can be modulated by a top-down influence from higher-order cognitive processes. This model allows us to further investigate both the precedence effect and the perceived-spatial-separation effect using laboratory animals.
(8) Frequency-following responses to pain vocalization in rats
In humans, brainstem frequency-following responses (FFRs), which are sustained potentials based on phase-locked neural activities, precisely preserve low-to-medium-frequency information of periodical sound waveforms. Interestingly, FFRs in humans can be modulated by both attention and language experience.
For the first time, we found that in rats FFRs to the audible component of vocal expressions of pain by conspecifics (rats’ “pain chatter”) can be recorded in not only the auditory midbrain, the inferior colliculus, but also the lateral nucleus of the amygdala. The rat’s FFRs to the pain chatter are remarkably resistant to noise masking.
When both the pain chatter and a masking noise is presented to each of the rat’s two ears, introducing a disparity between the interaural time difference for the chatter and that for the interaurally correlated noise markedly enhances the fundamental-frequency component of the FFRs to the chatter. A further increase of the binaural enhancement of the FFRs in the lateral nucleus of the amygdala can be produced by blocking excitatory glutamate transmissions in the auditory association cortex. Thus the amygdala, which is critical for auditory emotional perception, auditory emotional learning, and fear-conditioning modulation of auditory coding, preserves detailed acoustic features of species-specific emotional calls. This signal preservation is affected by both bottom-up binaural processing and top-down cortical modulations.
(9) Neural mechanisms underlying binaural inhibitory responses in the auditory midbrain
Both sound localization and the precedence effect largely depend on binaural integration in the central auditory system. The inferior colliculus occupies a critical position for processing binaural information. In mammals, the majority of binaural neurons in the inferior colliculus are so-called EI neurons, which receive excitatory inputs from the contralateral ear and inhibitory inputs from the ipsilateral ear. We investigated the neural mechanisms underlying the formation of binaural inhibition in the inferior colliculus and found that the dorsal nucleus of the lateral lemniscus, which is situated just below the inferior colliculus, is a key structure that shapes binaural inhibitory responses in the inferior colliculus by its inhibitory axonal projections to the inferior colliculus. Currently we are studying whether the binaural inhibitory mechanism is associated with the precedence effect and spatial unmasking.
(10) Animal models of schizophrenia
Human schizophrenics are characterized by both the deficit in the prefrontal-dysfunction-related flexibility and the reduction of prepulse inhibition. The neurodevelopmental hypothesis of schizophrenia emphasizes that certain early-life environmental factors have substantial influences upon the processes of brain maturation, and cause anatomical and functional abnormalities in the central nervous system. Accordingly, several animal models involving early-life manipulations are proposed. One of the early-life manipulations is isolation rearing after weaning.
We found that rats with selective lesions of the ventral medial prefrontal cortex, including the infralimbic area or prelimbic area, perform poorly in reversal learning but not in acquisition learning in a rotating T maze. Moreover, our recent results show that isolation-reared rats without clozapine injection perform significantly worse than socially-reared rats in reversal learning but not in acquisition learning. Chronic injection of clozapine in isolation-reared rats significantly improves reversal learning. This rodent reversal-learning model is useful not only for studying neural abnormalities associated with isolation rearing, particularly the neural mechanisms underlying schizophrenia, but also for developing new antipsychotics.
Moreover, we found that both prepulse inhibition and emotional-learning-induced enhancement of prepulse inhibition are much weaker in isolation-reared rats than those in socially-reared rats, and the prepulse inhibition enhancement depends on activity of metabotropic glutamate subtype 5 receptors. Because metabotropic glutamate subtype 5 receptors contribute to glutamatergic dysfunction observed in human patients of schizophrenia, we propose that we have advanced the prepulse-inhibition model for investigating both neural bases and cognitive features of schizophrenia.