Norm-Based Coding of Voice Identity in Human Auditory Cortex

Abstract

SummaryListeners exploit small interindividual variations around a generic acoustical structure to discriminate and identify individuals from their voice—a key requirement for social interactions. The human brain contains temporal voice areas (TVA) [1] involved in an acoustic-based representation of voice identity [2–6], but the underlying coding mechanisms remain unknown. Indirect evidence suggests that identity representation in these areas could rely on a norm-based coding mechanism [4, 7–11]. Here, we show by using fMRI that voice identity is coded in the TVA as a function of acoustical distance to two internal voice prototypes (one male, one female)—approximated here by averaging a large number of same-gender voices by using morphing [12]. Voices more distant from their prototype are perceived as more distinctive and elicit greater neuronal activity in voice-sensitive cortex than closer voices—a phenomenon not merely explained by neuronal adaptation [13, 14]. Moreover, explicit manipulations of distance-to-mean by morphing voices toward (or away from) their prototype elicit reduced (or enhanced) neuronal activity. These results indicate that voice-sensitive cortex integrates relevant acoustical features into a complex representation referenced to idealized male and female voice prototypes. More generally, they shed light on remarkable similarities in cerebral representations of facial and vocal identity.

Task-Dependent Decoding of Speaker and Vowel Identity from Auditory Cortical Response Patterns

“…First, the right middle STG/STS, as well as smaller subthreshold bilateral STG/STS and temporal plane clusters, showed stronger activation during the speaker task. This speaker task modulation confirms and extends previous reports of the involvement of these superior temporal regions in the passive and/or active processing of human voices (Belin et al, 2000;von Kriegstein et al, 2003;Andics et al, 2010;Moerel et al, 2012;Bonte et al, 2013;Latinus et al, 2013). Second, the right posterior STS/MTG showed stronger activation during the vowel task.…”

“…Beyond regional differences in task-specific activation levels, our multivariate decoding results demonstrate that distinct but overlapping response patterns across early and higher-order auditory cortex entail abstract, goal-dependent representations of individual speech stimuli; that is, the task dependency of speaker/ vowel decoding accuracies shows enhanced distinction of response patterns for individual speakers/vowels along the taskrelevant dimension. Speaker discrimination most consistently relied on voxels clustering in early auditory regions (HG/HS) and the temporal plane, as well as in regions along the middle to anterior (right) STG/STS, that overlap with the superior temporal voice areas (Belin et al, 2000;Moerel et al, 2012;Bonte et al, 2013;Latinus et al, 2013;Fig. 2B) and with right STG/STS regions recruited during voice recognition tasks (von Kriegstein et al, 2003;Lattner et al, 2005;Andics et al, 2010).…”

The effects of stimulus complexity on the preattentive processing of self-generated and nonself voices: An ERP study

“…We found no association between a voice's acoustic properties and the amplitudes of both the MMN and P3a to an SGV, yet earlier evidence had demonstrated that both F0 and formant frequencies are critical acoustic cues underlying successful voice identity recognition (e.g., Latinus et al, 2013;Xu et al, 2013). As expected, this lack of association was plausibly due to the fact that our ERP analysis controlled for the physical differences between the voice stimuli by using a Blike from like^subtraction approach.…”

“…Furthermore, recent studies have shown that listeners strongly rely on both perceived pitch (fundamental frequency: F0) and formant frequencies to extract voice identity information (Baumann & Belin, 2010;Belin, Bestelmeyer, Latinus, & Watson, 2011;Belin, Fecteau, & Bédard, 2004;Latinus, McAleer, Bestelmeyer, & Belin, 2013;Schweinberger, Kawahara, Simpson, Skuk, & Zäske, 2014;Schweinberger, Walther, Zäske, & Kovács, 2011;Xu et al, 2013). The existing evidence suggests that each speaker's voice is coded within temporal voice-sensitive regions as a function of its physical acoustic deviation regarding an internal vocal prototype (Latinus & Belin, 2012;Latinus et al, 2013;Schweinberger et al, 2011): More acoustically distant voices are not only perceived as being more Bdistinctive^ (Baumann & Belin, 2010;, but they also elicit increased activation within these temporal voice-sensitive regions (Latinus et al, 2013). Thus, considering the reported relationship between the physical features of the acoustic signal and the representation of speaker's identity, we probed whether the ERP correlates of automatic change detection and attention orienting to one's own voice were associated with the voice's acoustic properties.…”

Functional selectivity for face processing in the temporal voice area of early deaf individuals

“…The face-selective cross-modal recruitment of dTFA suggests that cross-modal effects do not occur uniformly across areas of the deaf cortex and supports the notion that cross-modal plasticity is related to the original functional specialization of the colonized brain regions (4,27). Indeed, temporal voice areas typically involved in an acoustic-based representation of voice identity (28) are shown here to code for facial identity discrimination ( Fig. 2A), which is in line with previous investigations in (Right) Suprathreshold (P = 0.05 FWE cluster-corrected over the whole brain) face-dependent PPI of right TVA in deaf subjects and significant differences between the deaf and the two control groups are superimposed on the MNI-ICBM152 template.…”