Neural signatures of disordered multi-talker speech perception in adults with normal hearing

Abstract

1In social settings, speech waveforms from nearby speakers mix together in our ear canals. The brain 2 unmixes the attended speech stream from the chorus of background speakers using a combination of fast 3 temporal processing and cognitive active listening mechanisms. Multi-talker speech perception is 4 vulnerable to aging or auditory abuse. We found that ~10% of adult visitors to our clinic have no 5 measurable hearing loss, yet offer a primary complaint of poor hearing. Multi-talker speech intelligibility 6 in these adults was strongly correlated with neural phase locking to frequency modulation (FM) cues, as 7 determined from ear canal EEG recordings. Combining neural temporal fine structure (TFS) processing 8 with pupil-indexed measures of cognitive listening effort could predict most of the individual variance in 9 speech intelligibility thresholds. These findings identify a confluence of disordered bottom-up and top-10 down processes that predict poor multi-talker speech perception and could be useful in next-generation 11 tests of hidden hearing disorders. 12 13 2015). Here, we apply parallel psychophysical and neurophysiological tests of sTFS processing in 1 combination with physiological measures of effortful listening to converge on a set of neural biomarkers 2 that identify poor multi-talker speech intelligibility in adults with clinically normal hearing. 3 4 Results 5Many individuals seek medical care for poor hearing but have no evidence of hearing loss 6We identified the first visit records of English-speaking adult patients from the Mass. Eye and Ear 7 audiology database over a 16-year period, with complete bilateral audiometric records at six octave 8 frequencies from 250 Hz to 8000 Hz according to the inclusion criteria in Figure 1A. Of the 106,787 patient 9 records that met these criteria, we found that approximately one out of every five individuals had no 10 clinical evidence of hearing loss, defined as thresholds > 20 dB HL at test frequencies up to 8 KHz (19,952, 11 19%, Figure 1B). The majority of these individuals were between 20-50 years old ( Figure 1C) and had no 12 conductive hearing impairment, nor focal threshold shifts or "notches" in their audiograms greater than 13 10 dB ( Fig. 1 -Fig. supplement 1A). The thresholds between their left and right ears were also symmetrical 14 within 10 dB for >95% of these patients ( Fig. 1 -Fig. supplement 1B). Despite these clinically normal 15measures of hearing, 45% of these individuals presented to the clinic reporting a primary complaint of 16 decreased hearing or hearing loss (Figure 1D). Absent any objective measure of hearing difficulty, these 17 patients are typically informed that their hearing is "normal" and that they are not expected to experience 18 communication problems. 19 20 Speech-in-noise intelligibility varies widely in individuals with clinically normal hearing 21

The role of cochlear place coding in the perception of frequency modulation

“…Modulations in frequency (FM) and amplitude (AM) carry critical information in biologically relevant sounds, such as speech, music, and animal vocalizations (Attias and Schreiner, 1997;Nelken et al, 1999). In humans, AM is crucial for understanding speech in quiet (Shannon et al, 1995;Smith et al, 2002), while FM is particularly important for perceiving melodies, recognizing talkers, determining speech prosody and emotion, and may aid in the perception of speech presented in competing background sounds (Zeng et al, 2005;Strelcyk and Dau, 2009;Sheft et al, 2012;Johannesen et al, 2016;Lopez-Poveda et al, 2017;Parthasarathy et al, 2019).…”

“…Modulations in frequency (FM) and amplitude (AM) carry critical information in biologically relevant sounds, such as speech, music, and animal vocalizations ( Attias and Schreiner, 1997 ; Nelken et al, 1999 ). In humans, AM is crucial for understanding speech in quiet ( Shannon et al, 1995 ; Smith et al, 2002 ), while FM is particularly important for perceiving melodies, recognizing talkers, determining speech prosody and emotion, and may aid in the perception of speech presented in competing background sounds ( Zeng et al, 2005 ; Strelcyk and Dau, 2009 ; Sheft et al, 2012 ; Johannesen et al, 2016 ; Lopez-Poveda et al, 2017 ; Parthasarathy et al, 2019 ). The perception of FM at both slow and fast modulation rates is often degraded in older people and those with hearing loss ( Lacher-Fougère and Demany, 1998 ; Moore and Skrodzka, 2002 ; He et al, 2007 ; Strelcyk and Dau, 2009 ; Grose and Mamo, 2012 ; Paraouty et al, 2016 ; Wallaert et al, 2016 ; Paraouty and Lorenzi, 2017 ; Whiteford et al, 2017 ).…”

“…This discrepancy is important, because it is FM at low frequencies and slow modulation rates that is most critical for human communication, including speech and music, as well as many animal vocalizations ( Attias and Schreiner, 1997 ; Nelken et al, 1999 ). The enhanced sensitivity to slow FM at low carrier frequencies has been explained in terms of an additional neural code based on stimulus-driven spike timing in the auditory nerve that is phase-locked to the temporal fine structure of the stimulus ( Rose et al, 1967 ; Moore and Sek, 1995 ; Parthasarathy et al, 2019 ). Although a code based on time intervals between phase-locked neural spikes can potentially provide greater accuracy ( Siebert, 1970 ; Heinz et al, 2001 ), be maintained to some extent in the auditory brainstem ( Paraouty et al, 2018 ), and be used for spatial localization ( Moiseff and Konishi, 1981 ; Grothe et al, 2010 ), it is not known whether or how this timing information is extracted by higher stages of the auditory system to encode periodicity (pitch) and FM.…”

Temporal integration of monaural and dichotic frequency modulation

“…Either the instantaneous amplitude of the signal is encoded by phase-locked firing (i.e., TFS of the stimulus) or FM is converted to amplitude modulation (AM) by changes in the output of the auditory filters fixed in place along the cochlear partition ( Zwicker, 1956 ; Khanna and Teich, 1989 ; Sęk and Moore, 1995 ; Moore and Sęk, 1996 ). Differences in the relationship between the modulation frequency and detection for FM and AM signals have previously been interpreted as evidence for the importance of TFS cues for FM detection at low modulation frequencies ( Rose et al , 1967 ; Moore and Sęk, 1995 ; Parthasarathy et al , 2019 ). However, there is converging evidence that place cues alone may be able to explain FM detection even at low modulation rates.…”

A biologically oriented algorithm for spatial sound segregation

“…Attending to a single conversation partner in the presence of multiple distracting talkers (i.e., the Cocktail Party Problem, CPP) is a complicated and difficult task for machines and humans [1][2][3]. While some normal-hearing listeners can accomplish this task with relative ease, other groups of listeners report great difficulty -such as those with sensorineural hearing loss [4][5][6], cochlear implant users [7][8][9][10], subgroups of children [11] and adults with "hidden hearing loss" [12][13][14]. At a cocktail party, talkers are distributed in space, and normal-hearing listeners appear to make use of spatial cues (i.e., interaural timing and level differences, or ITDs and ILDs, respectively) to perceptually localize and segregate sound mixtures into individual spatial components.…”