skip to main content

This content will become publicly available on January 21, 2023

Title: Parallel processing in speech perception with local and global representations of linguistic context
Speech processing is highly incremental. It is widely accepted that human listeners continuously use the linguistic context to anticipate upcoming concepts, words, and phonemes. However, previous evidence supports two seemingly contradictory models of how a predictive context is integrated with the bottom-up sensory input: Classic psycholinguistic paradigms suggest a two-stage process, in which acoustic input initially leads to local, context-independent representations, which are then quickly integrated with contextual constraints. This contrasts with the view that the brain constructs a single coherent, unified interpretation of the input, which fully integrates available information across representational hierarchies, and thus uses contextual constraints to modulate even the earliest sensory representations. To distinguish these hypotheses, we tested magnetoencephalography responses to continuous narrative speech for signatures of local and unified predictive models. Results provide evidence that listeners employ both types of models in parallel. Two local context models uniquely predict some part of early neural responses, one based on sublexical phoneme sequences, and one based on the phonemes in the current word alone; at the same time, even early responses to phonemes also reflect a unified model that incorporates sentence-level constraints to predict upcoming phonemes. Neural source localization places the anatomical origins of the different predictive more » models in nonidentical parts of the superior temporal lobes bilaterally, with the right hemisphere showing a relative preference for more local models. These results suggest that speech processing recruits both local and unified predictive models in parallel, reconciling previous disparate findings. Parallel models might make the perceptual system more robust, facilitate processing of unexpected inputs, and serve a function in language acquisition. « less
Authors:
; ; ; ; ;
Award ID(s):
1734892
Publication Date:
NSF-PAR ID:
10321962
Journal Name:
eLife
Volume:
11
ISSN:
2050-084X
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract
    Speech processing is highly incremental. It is widely accepted that human listeners continuously use the linguistic context to anticipate upcoming concepts, words, and phonemes. However, previous evidence supports two seemingly contradictory models of how a predictive context is integrated with the bottom-up sensory input: Classic psycholinguistic paradigms suggest a two-stage process, in which acoustic input initially leads to local, context-independent representations, which are then quickly integrated with contextual constraints. This contrasts with the view that the brain constructs a single coherent, unified interpretation of the input, which fully integrates available information across representational hierarchies, and thus uses contextual constraints to modulate even the earliest sensory representations. To distinguish these hypotheses, we tested magnetoencephalography responses to continuous narrative speech for signatures of local and unified predictive models. Results provide evidence that listeners employ both types of models in parallel. Two local context models uniquely predict some part of early neural responses, one based on sublexical phoneme sequences, and one based on the phonemes in the current word alone; at the same time, even early responses to phonemes also reflect a unified model that incorporates sentence-level constraints to predict upcoming phonemes. Neural source localization places the anatomical origins of the differentMore>>
  2. Abstract

    Partial speech input is often understood to trigger rapid and automatic activation of successively higher-level representations of words, from sound to meaning. Here we show evidence from magnetoencephalography that this type of incremental processing is limited when words are heard in isolation as compared to continuous speech. This suggests a less unified and automatic word recognition process than is often assumed. We present evidence from isolated words that neural effects of phoneme probability, quantified by phoneme surprisal, are significantly stronger than (statistically null) effects of phoneme-by-phoneme lexical uncertainty, quantified by cohort entropy. In contrast, we find robust effects of both cohort entropy and phoneme surprisal during perception of connected speech, with a significant interaction between the contexts. This dissociation rules out models of word recognition in which phoneme surprisal and cohort entropy are common indicators of a uniform process, even though these closely related information-theoretic measures both arise from the probability distribution of wordforms consistent with the input. We propose that phoneme surprisal effects reflect automatic access of a lower level of representation of the auditory input (e.g., wordforms) while the occurrence of cohort entropy effects is task sensitive, driven by a competition process or a higher-level representation thatmore »is engaged late (or not at all) during the processing of single words.

    « less
  3. Cerebellar granule cells (GrCs) are usually regarded as a uniform cell type that collectively expands the coding space of the cerebellum by integrating diverse combinations of mossy fiber inputs. Accordingly, stable molecularly or physiologically defined GrC subtypes within a single cerebellar region have not been reported. The only known cellular property that distinguishes otherwise homogeneous GrCs is the correspondence between GrC birth timing and the depth of the molecular layer to which their axons project. To determine the role birth timing plays in GrC wiring and function, we developed genetic strategies to access early- and late-born GrCs. We initiated retrograde monosynaptic rabies virus tracing from control (birth timing unrestricted), early-born, and late-born GrCs, revealing the different patterns of mossy fiber input to GrCs in vermis lobule 6 and simplex, as well as to early- and late-born GrCs of vermis lobule 6: sensory and motor nuclei provide more input to early-born GrCs, while basal pontine and cerebellar nuclei provide more input to late-born GrCs. In vivo multidepth two-photon Ca 2+ imaging of axons of early- and late-born GrCs revealed representations of diverse task variables and stimuli by both populations, with modest differences in the proportions encoding movement, reward anticipation, and rewardmore »consumption. Our results suggest neither organized parallel processing nor completely random organization of mossy fiber→GrC circuitry but instead a moderate influence of birth timing on GrC wiring and encoding. Our imaging data also provide evidence that GrCs can represent generalized responses to aversive stimuli, in addition to recently described reward representations.« less
  4. Aging is associated with an exaggerated representation of the speech envelope in auditory cortex. The relationship between this age-related exaggerated response and a listener’s ability to understand speech in noise remains an open question. Here, information-theory-based analysis methods are applied to magnetoencephalography recordings of human listeners, investigating their cortical responses to continuous speech, using the novel nonlinear measure of phase-locked mutual information between the speech stimuli and cortical responses. The cortex of older listeners shows an exaggerated level of mutual information, compared with younger listeners, for both attended and unattended speakers. The mutual information peaks for several distinct latencies: early (∼50 ms), middle (∼100 ms), and late (∼200 ms). For the late component, the neural enhancement of attended over unattended speech is affected by stimulus signal-to-noise ratio, but the direction of this dependency is reversed by aging. Critically, in older listeners and for the same late component, greater cortical exaggeration is correlated with decreased behavioral inhibitory control. This negative correlation also carries over to speech intelligibility in noise, where greater cortical exaggeration in older listeners is correlated with worse speech intelligibility scores. Finally, an age-related lateralization difference is also seen for the ∼100 ms latency peaks, where older listeners showmore »a bilateral response compared with younger listeners’ right lateralization. Thus, this information-theory-based analysis provides new, and less coarse-grained, results regarding age-related change in auditory cortical speech processing, and its correlation with cognitive measures, compared with related linear measures. NEW & NOTEWORTHY Cortical representations of natural speech are investigated using a novel nonlinear approach based on mutual information. Cortical responses, phase-locked to the speech envelope, show an exaggerated level of mutual information associated with aging, appearing at several distinct latencies (∼50, ∼100, and ∼200 ms). Critically, for older listeners only, the ∼200 ms latency response components are correlated with specific behavioral measures, including behavioral inhibition and speech comprehension.« less
  5. Recently, deep feedforward neural networks have achieved considerable success in modeling biological sensory processing, in terms of reproducing the input-output map of sensory neurons. However, such models raise profound questions about the very nature of explanation in neuroscience. Are we simply replacing one complex system (a biological circuit) with another (a deep network), without understanding either? Moreover, beyond neural representations, are the deep network's computational mechanisms for generating neural responses the same as those in the brain? Without a systematic approach to extracting and understanding computational mechanisms from deep neural network models, it can be difficult both to assess the degree of utility of deep learning approaches in neuroscience, and to extract experimentally testable hypotheses from deep networks. We develop such a systematic approach by combining dimensionality reduction and modern attribution methods for determining the relative importance of interneurons for specific visual computations. We apply this approach to deep network models of the retina, revealing a conceptual understanding of how the retina acts as a predictive feature extractor that signals deviations from expectations for diverse spatiotemporal stimuli. For each stimulus, our extracted computational mechanisms are consistent with prior scientific literature, and in one case yields a new mechanistic hypothesis. Thusmore »overall, this work not only yields insights into the computational mechanisms underlying the striking predictive capabilities of the retina, but also places the framework of deep networks as neuroscientific models on firmer theoretical foundations, by providing a new roadmap to go beyond comparing neural representations to extracting and understand computational mechanisms.« less