Investigating the effect of spatial separation on the detection of sounds in competition, by examining electrophysiological responses from the brainstem and auditory cortex
AffiliationAudiology and Speech Pathology
Centre for Neuroscience
Document TypePhD thesis
Access StatusOpen Access
© 2016 Dr. Nematollah Rouhbakhsh
Human communication frequently takes place in noisy environments. In these environments, successful understanding of speech is dependent on an individual’s ability to extract and use spatial cues for separating speech from distracting noise. When speech and noise are separated spatially, the speech reception threshold (SRT) is reduced, and this is referred to in the literature as spatial release from masking (SRM). SRM is in most part due to acoustic cues arising from the differences in time and intensity of signals arriving at each of our two ears (i.e. interaural time (ITD) and interaural level (ILD) differences). ITDs and ILDs have in general been investigated through psychoacoustic studies. However, the electrophysiological correlates of these acoustic cues have only been investigated individually. For this reason, a novel experiment was designed to investigate the effect of spatial separation on the detection of target sound in competition with distractor stimuli in a more realistic experimental environment in which both ITD and ILD cues were present. The primary aim of this thesis was to determine whether it is possible to identify a neural representation of SRM in the electrophysiological responses recorded from either the brainstem or auditory cortex, or both, using experimental stimuli conveying ILD and ITD cues. This research was conducted in two primary studies. The first study investigated whether the frequency-following response (FFR) in response to the fundamental frequency (F0) of a speech sound could be used to demonstrate SRM at different signal-to-noise ratios (SNRs), and what role of attentional mechanisms might play in spatial processing. FFRs were recorded in eighteen normally hearing participants. Participants were presented through headphones with a synthesized steady-state vowel /u/ with an F0 of 110 Hz and a 250 ms duration at 60 dB SPL. This vowel was labelled as the target stimulus. To be able to measure the effects of attention, a deviant stimulus was interspersed randomly throughout the target stimuli. It was presented 5% of the time, at 52 dB SPL. The role of attention in SRM was measured in two phases. In the “attended” phase, participants were asked to count the number of deviants that occurred. It was assumed that, while identifying the number of deviant stimuli, the participant was actively listening to the stimuli. In the “non-attended” phase, participants were asked to ignore both the target and deviant stimuli, and any distractors. The distractors were two continuous different stories spoken by the same speaker. Target, deviant and distractor stimuli were convolved with head-related transfer functions (HRTFs) to create two spatial conditions: the co-located condition with targets, deviants and distractors coming from 0⁰ azimuth; and the separated condition with the targets and deviants at 0⁰, but with each distractor shifted to each side (± 90⁰ azimuth). Three SNRs were considered (-5, -0, and 5 dB). The amplitudes of the FFR in response to F0 were determined and analysed. The results of study 1 revealed a significant effect of spatial separation. The effect of spatial separation was found only at the lower SNR. Spatially separating maskers from the target stimuli resulted in a significant larger amplitude of the FFR in response to the target F0. The spatial advantage obtained objectively was equivalent to an SNR increase of 3.3 dB. A significant effect for attention was found when participants actively focused on the target, as demonstrated by larger FFR amplitudes. However, no significant interactions were found between spatial separation and the level of attention. The findings of the first study suggest that binaural processing relevant to SRM may be reflected by phase locked neural activity in the brainstem. However, this objective measure may only be noticeable in relatively noisy environments. Furthermore, SRM may start early in the central auditory pathways regardless of one attending to the target stimuli or not. This last observation means that - although this thesis focuses on adults - an extrapolation potentially could be made towards the use with younger individuals, however with consideration of their brain differences with adults and the AEPs evoked from those brains. The lack of dependence on attention might be beneficial in investigating SRM in this population, where it is difficult to keep attention and one has to rely on objective techniques that do not require attending to the target stimulus. Conversely, the lack of interaction with attention may mean that the mechanism responsible for the objective results may be different from the mechanism primarily responsible for SRM. To identify whether objective markers of SRM can be recorded in either the brainstem or cortex (or both), a second study was conducted. In the second study, auditory brainstem responses (ABRs), FFRs, and cortical auditory evoked potentials (CAEPs) were recorded simultaneously from thirteen normally hearing adults in response to 200 target stimulus blocks. Each target stimulus block comprised of a series of 11 tone complexes (TCs), with each TC having a specific F0 and a duration of 30 ms, separated by a 30 ms interstimulus interval (ISI), resulting in a target stimulus block with a total duration of 630 ms. The blocks were repeated every 1200 ms. Two different target stimulus block paradigms were considered; flat and staircase. The ‘flat’ blocks had TCs with a constant fundamental frequency F0 of 325 Hz (and harmonics up to 6 kHz). In the ‘staircase’ blocks, the fundamental frequency of each TCs was reduced in steps of 30 Hz from 475 to 175 Hz, again with harmonics up to 6 kHz. ABRs were recorded to the onsets of the 30-ms TCs. FFRs were recorded in response to the F0s of the TCs, and CAEPs to the onsets of the target stimulus blocks. The distractor blocks, in contrast, comprised of blocks of TCs that were similar in number and duration to the target stimuli, but randomized in their F0 distribution from 100 to 550 Hz and jittered in time (+/- 15 ms) around the onset of the target TC. Both target and distractor stimuli were convolved with head-related transfer functions (HRTFs) and presented under headphones. The target stimuli were presented at 0⁰ azimuth. The distractors were co-located (at 0⁰ azimuth) and spatially separated (at ±90⁰ azimuth) from the targets. The targets were presented at SNRs of -5, 0, 5, 10 and 15 dB SNR, and at 60 dB SPL. After extraction and analysis of ABR amplitudes and latencies, and FFR amplitudes, the results of the second study revealed a significant effect of SRM as seen in a decrease in ABR latency for both flat and staircase target stimuli when spatially separating maskers from the target. FFR amplitude (only measured with the flat stimuli) was significantly larger in the separated condition, and a significant decrease in CAEP latencies (for the staircase stimuli) was found, but only at the lowest tested SNR of -5 dB. These results, particularly the FFR, confirmed the results obtained in the first study, i.e. separating distractors from the target, regardless of the type of stimulus being used, resulted in enhancing FFR F0 amplitude. However, due to noisy data, the observations at the cortical level need to be confirmed in a follow-up study. The spatial advantage was equivalent to a SNR increase of 4.3 dB for FFR amplitude (for the flat stimuli), and 13.8 dB for ABR latency, 11.2 dB for CAEP P1 and 19.9 dB for CAEP N1 latencies (for the staircase stimuli). The findings of the second study suggest that it is possible objectively to record SRM in both the brainstem and auditory cortex simultaneously at lower SNRs. This suggests that the central auditory system is able to squelch background noise via processing of spatial information, and that this capacity is higher in more challenging listening environments. Taken together, the results from the first and second studies suggest that it is feasible to use electrophysiological measures as a means of investigating the central auditory mechanisms, which contribute to SRM in the brainstem and cortex simultaneously. It is speculated that SRM occurs mainly at the level of the brainstem and is present at -5 dB SNR (i.e. difficult listening environments). The finding that SRM was primarily at lower SNRs is in reality not a clinical concern, as lower SNRs represent the environments in which SRM is generally found to be beneficial for the listener. Potential applications may be found in developing an objective detection test for spatial processing disorder (SPD), a condition in which normal-hearing individuals are unable to exploit the binaural mechanism of SRM when listening in noisy environments, i.e., a deficiency in selectively attending to target sounds, which are not spatially co-located with distractor sounds. Further studies are needed to investigate the effects of attention and SRM on brainstem and cortical responses in different populations including children, elderly, and people with SPD.
Keywordsspatial processing; spatial release from masking; spatial advantage; auditory evoked potentials; frequency-following response; auditory brainstem response; cortical auditory evoked response
- Click on "Export Reference in RIS Format" and choose "open with... Endnote".
- Click on "Export Reference in RIS Format". Login to Refworks, go to References => Import References