Predictive Brain Mechanisms in Sound-to-Meaning Mapping during Speech Processing

Common and unique neural mechanisms of social and nonsocial conflict resolving and adaptation

Humans often need to deal with various forms of information conflicts that arise when they receive inconsistent information. However, it remains unclear how we resolve them and whether the brain may recruit similar or distinct brain mechanisms to process different domains (e.g. social vs. nonsocial) of conflicts. To address this, we used functional magnetic resonance imaging and scanned 50 healthy participants when they were asked to perform 2 Stroop tasks with different forms of conflicts: social (i.e. face-gender incongruency) and nonsocial (i.e. color-word incongruency) conflicts. Neuroimaging results revealed that the ventral lateral prefrontal cortex was generally activated in processing incongruent versus congruent stimuli regardless of the task type, serving as a common mechanism for conflict resolving across domains. Notably, trial-based and model-based results jointly demonstrated that the dorsal and rostral medial prefrontal cortices were uniquely engaged in processing social incongruent stimuli, suggesting distinct neural substrates of social conflict resolving and adaptation. The findings uncover that the common but unique brain mechanisms are recruited when humans resolve and adapt to social conflicts.

Predicting speech from a cortical hierarchy of event-based time scales

How do predictions in the brain incorporate the temporal unfolding of context in our natural environment? We here provide evidence for a neural coding scheme that sparsely updates contextual representations at the boundary of events. This yields a hierarchical, multilayered organization of predictive language comprehension. Training artificial neural networks to predict the next word in a story at five stacked time scales and then using model-based functional magnetic resonance imaging, we observe an event-based “surprisal hierarchy” evolving along a temporoparietal pathway. Along this hierarchy, surprisal at any given time scale gated bottom-up and top-down connectivity to neighboring time scales. In contrast, surprisal derived from continuously updated context influenced temporoparietal activity only at short time scales. Representing context in the form of increasingly coarse events constitutes a network architecture for making predictions that is both computationally efficient and contextually diverse.

Predicting speech from a cortical hierarchy of event-based time scales

How do predictions in the brain incorporate the temporal unfolding of context in our natural environment? We here provide evidence for a neural coding scheme that sparsely updates contextual representations at the boundary of events. This yields a hierarchical, multilayered organization of predictive language comprehension. Training artificial neural networks to predict the next word in a story at five stacked time scales and then using model-based functional magnetic resonance imaging, we observe an event-based “surprisal hierarchy” evolving along a temporoparietal pathway. Along this hierarchy, surprisal at any given time scale gated bottom-up and top-down connectivity to neighboring time scales. In contrast, surprisal derived from continuously updated context influenced temporoparietal activity only at short time scales. Representing context in the form of increasingly coarse events constitutes a network architecture for making predictions that is both computationally efficient and contextually diverse.

Listening to speech with a guinea pig-to-human brain-to-brain interface

Nicolelis wrote in his 2003 review on brain-machine interfaces (BMIs) that the design of a successful BMI relies on general physiological principles describing how neuronal signals are encoded. Our study explored whether neural information exchanged between brains of different species is possible, similar to the information exchange between computers. We show for the first time that single words processed by the guinea pig auditory system are intelligible to humans who receive the processed information via a cochlear implant. We recorded the neural response patterns to single-spoken words with multi-channel electrodes from the guinea inferior colliculus. The recordings served as a blueprint for trains of biphasic, charge-balanced electrical pulses, which a cochlear implant delivered to the cochlear implant user’s ear. Study participants completed a four-word forced-choice test and identified the correct word in 34.8% of trials. The participants' recognition, defined by the ability to choose the same word twice, whether right or wrong, was 53.6%. For all sessions, the participants received no training and no feedback. The results show that lexical information can be transmitted from an animal to a human auditory system. In the discussion, we will contemplate how learning from the animals might help developing novel coding strategies.

The Cortical Maps of Hierarchical Linguistic Structures during Speech Perception

The hierarchical nature of language requires human brain to internally parse connected-speech and incrementally construct abstract linguistic structures. Recent research revealed multiple neural processing timescales underlying grammar-based configuration of linguistic hierarchies. However, little is known about where in the whole cerebral cortex such temporally scaled neural processes occur. This study used novel magnetoencephalography source imaging techniques combined with a unique language stimulation paradigm to segregate cortical maps synchronized to 3 levels of linguistic units (i.e., words, phrases, and sentences). Notably, distinct ensembles of cortical loci were identified to feature structures at different levels. The superior temporal gyrus was found to be involved in processing all 3 linguistic levels while distinct ensembles of other brain regions were recruited to encode each linguistic level. Neural activities in the right motor cortex only followed the rhythm of monosyllabic words which have clear acoustic boundaries, whereas the left anterior temporal lobe and the left inferior frontal gyrus were selectively recruited in processing phrases or sentences. Our results ground a multi-timescale hierarchical neural processing of speech in neuroanatomical reality with specific sets of cortices responsible for different levels of linguistic units.

Spatiotemporal dynamics of predictive brain mechanisms during speech processing: an MEG study

Rapid and efficient speech processing benefits from the prediction derived from prior expectations based on the identification of individual words. It is known that speech processing is carried out within a distributed frontotemporal network. However, the spatiotemporal causal dynamics of predictive brain mechanisms in sound-to-meaning mapping within this network remain unclear. Using magnetoencephalography, we adopted a semantic anomaly paradigm which consists of expected, unexpected and time-reversed Mandarin Chinese speech, and localized the effects of violated expectation in frontotemporal brain regions, the sensorimotor cortex and the supramarginal gyrus from 250 ms relative to the target words. By further investigating the causal cortical dynamics, we provided the description of the causal dynamic network within the framework of the dual stream model, and highlighted the importance of the connections within the ventral pathway, the top-down modulation from the left inferior frontal gyrus and the cross-stream integration during the speech processing of violated expectation.

Understanding rostral-caudal auditory cortex contributions to auditory perception

There are functional and anatomical distinctions between the neural systems involved in the recognition of sounds in the environment and those involved in the sensorimotor guidance of sound production and the spatial processing of sound. Evidence for the separation of these processes has historically come from disparate literatures on the perception and production of speech, music and other sounds. More recent evidence indicates that there are computational distinctions between the rostral and caudal primate auditory cortex that may underlie functional differences in auditory processing. These functional differences may originate from differences in the response times and temporal profiles of neurons in the rostral and caudal auditory cortex, suggesting that computational accounts of primate auditory pathways should focus on the implications of these temporal response differences.

Aging-related differences in the cortical network subserving intelligible speech

Language communication is crucial throughout the lifespan. The current study investigated how aging affects the brain network subserving intelligible speech. Using functional magnetic resonance imaging, we compared brain responses to intelligible and unintelligible speech between older and young adults. Univariate and multivariate analyses revealed reduced brain activation and lower regional pattern distinctions in response to intelligible versus unintelligible speech in the left anterior superior temporal gyrus (aSTG) and the left inferior frontal gyrus (IFG) in the older compared with young adults. Notably, the functional connectivity between the left IFG and the left angular gyrus (AG) was increased and a significantly enhanced bidirectional effective connectivity between the left aSTG and the left AG was observed in the older adults for processing speech intelligibility. Our study revealed aging-related differences in the cortical activity for intelligible speech and suggested that increased frontal-temporal-parietal functional integration may help facilitate spoken language processing in older adults.

Is there a prediction network? Meta-analytic evidence for a cortical-subcortical network likely subserving prediction

Predictive coding is an increasingly influential and ambitious concept in neuroscience viewing the brain as a 'hypothesis testing machine' that constantly strives to minimize prediction error, the gap between its predictions and the actual sensory input. Despite the invaluable contribution of this framework to the formulation of brain function, its neuroanatomical foundations have not been fully defined. To address this gap, we conducted activation likelihood estimation (ALE) meta-analysis of 39 neuroimaging studies of three functional domains (action perception, language and music) inherently involving prediction. The ALE analysis revealed a widely distributed brain network encompassing regions within the inferior and middle frontal gyri, anterior insula, premotor cortex, pre-supplementary motor area, temporoparietal junction, striatum, thalamus/subthalamus and the cerebellum. This network is proposed to subserve domain-general prediction and its relevance to motor control, attention, implicit learning and social cognition is discussed in light of the predictive coding scheme. Better understanding of the presented network may help advance treatments of neuropsychiatric conditions related to aberrant prediction processing and promote cognitive enhancement in healthy individuals.

Reanalyzing neurocognitive data on the role of the motor system in speech perception within COSMO, a Bayesian perceptuo-motor model of speech communication

While neurocognitive data provide clear evidence for the involvement of the motor system in speech perception, its precise role and the way motor information is involved in perceptual decision remain unclear. In this paper, we discuss some recent experimental results in light of COSMO, a Bayesian perceptuo-motor model of speech communication. COSMO enables us to model both speech perception and speech production with probability distributions relating phonological units with sensory and motor variables. Speech perception is conceived as a sensory-motor architecture combining an auditory and a motor decoder thanks to a Bayesian fusion process. We propose the sketch of a neuroanatomical architecture for COSMO, and we capitalize on properties of the auditory vs. motor decoders to address three neurocognitive studies of the literature. Altogether, this computational study reinforces functional arguments supporting the role of a motor decoding branch in the speech perception process.

Focal versus distributed temporal cortex activity for speech sound category assignment

When listening to speech, phonemes are represented in a distributed fashion in our temporal and prefrontal cortices. How these representations are selected in a phonemic decision context, and in particular whether distributed or focal neural information is required for explicit phoneme recognition, is unclear. We hypothesized that focal and early neural encoding of acoustic signals is sufficiently informative to access speech sound representations and permit phoneme recognition. We tested this hypothesis by combining a simple speech-phoneme categorization task with univariate and multivariate analyses of fMRI, magnetoencephalography, intracortical, and clinical data. We show that neural information available focally in the temporal cortex prior to decision-related neural activity is specific enough to account for human phonemic identification.

Percepts and words can be decoded from distributed neural activity measures. However, the existence of widespread representations might conflict with the more classical notions of hierarchical processing and efficient coding, which are especially relevant in speech processing. Using fMRI and magnetoencephalography during syllable identification, we show that sensory and decisional activity colocalize to a restricted part of the posterior superior temporal gyrus (pSTG). Next, using intracortical recordings, we demonstrate that early and focal neural activity in this region distinguishes correct from incorrect decisions and can be machine-decoded to classify syllables. Crucially, significant machine decoding was possible from neuronal activity sampled across different regions of the temporal and frontal lobes, despite weak or absent sensory or decision-related responses. These findings show that speech-sound categorization relies on an efficient readout of focal pSTG neural activity, while more distributed activity patterns, although classifiable by machine learning, instead reflect collateral processes of sensory perception and decision.