Biology subjects

Chanoine, V.

Publications and source records attributed to Chanoine, V..

5 recordsLinked to original sources

bioRxiv2024

From movements to words: action monitoring in the medial frontal cortex along a caudal to rostral prediction error gradient

Speech error monitoring recruits the medial frontal cortex (MFC) region in the human brain. Error monitoring-related activity in the MFC has been interpreted both in terms of conflict monitoring and feedback-driven control, but as similar regions of the MFC are implicated in various levels of behavioral control ranging from basic motor movement control to high-level cognitive control functions, a more comprehensive account is needed. Moreover, as speech errors and other actions that involve varying control demands engage a widespread yet partially overlapping set of regions of the MFC, such an account should ideally explain the anatomical distribution of error-related functional activations within the MFC. Here we wanted to assess the hypothesis that the MFC has a similar role in the evaluation of action outcomes for motor and mental actions, operating along a rostral-caudal gradient of higher-lower level control demands involving prediction errors from both sensory and epistemic sources. To this end, we conducted an individual-specific annotation of task-fMRI BOLD activation peaks related to speech errors versus correct productions (i.e. that involve the largest cognitive control demands, Study I and II), tongue movement monitoring (i.e. that involve an intermediate level of cognitive and motor control demands) and tongue movement (i.e. that involve only motor control demands, Study II) in the MFC region. Results revealed overlapping clusters across the three contrasts across the MFC, but importantly both the number of peaks and their relative position along the rostral caudal axis were consistent with a hierarchical rostral caudal processing gradient in the MFC. While tongue movement showed more caudal activation in the MFC, speech errors showed more rostral activation, and tongue movement monitoring patterned in between. Furthermore, the combined results of both studies suggested that activation peaks were located more dorsally for participants that had a paracingulate gyrus, replicating a previously documented effect for movement and further supporting a common functional role of the MFC across very distinct actions.

neuroscience↗

bioRxiv2024

Revealing the co-existence of written and spoken language coding neural populations in the left-ventral occipitotemporal cortex

Reading relies on the ability to map written symbols with speech sounds. The left ventral occipitotemporal cortex (left-vOT) plays a crucial role in this process. Through the automatization of the mapping ability, this specific part of the ventral visual pathway (a.k.a., the Visual Word Form Area) progressively becomes specialized in written word recognition. Yet, despite its key role in reading, the area also responds to speech. This observation raises questions about the actual nature of neural representations encoded in the left-vOT and, therefore, the underlying mechanism of the cross-modal responses. Here, we addressed this issue by applying fine-grained analyses of within- and cross-modal repetition suppression effects (RSEs) and Multi-Voxel Pattern Analyses in fMRI and sEEG experiments. Convergent evidence across analysis methods and protocols showed significant RSEs and successful decoding in both within-modal visual and auditory conditions suggesting that subpopulations of neurons within the left-vOT distinctly encode written and spoken language inputs. This functional organization of neural populations enables the area to respond to speech input directly and indirectly, i.e., after speech sounds are converted to orthographic representations. The finding opens further discussions on how the human brain may be prepared and adapted for an acquisition of a complex ability such as reading. Significance StatementLearning to read generates new functional responses in neurons in the left ventral visual pathway. Soon after reading acquisition, these neurons become specialized in processing known scripts, thus leading to the functional designation of the "Visual Word Form Area" (VWFA). However, controversies remain regarding the nature of neural representations encoded in this "reading" region, as its activation to speech is also reported. We investigate the neural mechanism(s) underlying these bimodal responses using within and cross-modal repetition suppression and decoding protocols. fMRI and sEEG experiments provided converging evidence indicating that, despite its specialization in reading, VWFA also contained subpopulations of neurons that encode speech. This functional organization could reveal why neurons at this anatomical location are ideal for reading acquisition.

neuroscience↗

bioRxiv2023

Dimensionality and ramping: Signatures of sentence integration in the dynamics of brains and deep language models

A sentence is more than the sum of its words: its meaning depends on how they combine with one another. The brain mechanisms underlying such semantic composition remain poorly understood. To shed light on the neural vector code underlying semantic composition, we introduce two hypotheses: First, the intrinsic dimensionality of the space of neural representations should increase as a sentence unfolds, paralleling the growing complexity of its semantic representation, and second, this progressive integration should be reflected in ramping and sentence-final signals. To test these predictions, we designed a dataset of closely matched normal and Jabberwocky sentences (composed of meaningless pseudo words) and displayed them to deep language models and to 11 human participants (5 men and 6 women) monitored with simultaneous magneto-encephalography and intracranial electro-encephalography. In both deep language models and electrophysiological data, we found that representational dimensionality was higher for meaningful sentences than Jabberwocky. Furthermore, multivariate decoding of normal versus Jabberwocky confirmed three dynamic patterns: (i) a phasic pattern following each word, peaking in temporal and parietal areas, (ii) a ramping pattern, characteristic of bilateral inferior and middle frontal gyri, and (iii) a sentence-final pattern in left superior frontal gyrus and right orbitofrontal cortex. These results provide a first glimpse into the neural geometry of semantic integration and constrain the search for a neural code of linguistic composition. Significance statementStarting from general linguistic concepts, we make two sets of predictions in neural signals evoked by reading multi-word sentences. First, the intrinsic dimensionality of the representation should grow with additional meaningful words. Second, the neural dynamics should exhibit signatures of encoding, maintaining, and resolving semantic composition. We successfully validated these hypotheses in deep Neural Language Models, artificial neural networks trained on text and performing very well on many Natural Language Processing tasks. Then, using a unique combination of magnetoencephalography and intracranial electrodes, we recorded high-resolution brain data from human participants while they read a controlled set of sentences. Time-resolved dimensionality analysis showed increasing dimensionality with meaning, and multivariate decoding allowed us to isolate the three dynamical patterns we had hypothesized.

neuroscience↗

bioRxiv2022

Graph theoretical analysis reveals the adaptive role of the left ventral occipito-temporal cortex in the brain networks during speech processing

The left ventral occipito-temporal cortex (left-vOT) plays a key role in reading. Several studies have also reported its activation during speech processing, suggesting that it may play a role beyond written word recognition. Here, we adopt a graph theoretical analysis to investigate the functional role of this area in the whole-brain network while participants processed spoken sentences in different tasks. We find that its role and interactions with other areas changes in an adaptive manner. In a low-level speech perception task, the left-vOT is part of the visual network and acts as a connector that supports the communication with other cognitive systems. When speech comprehension is required, the area becomes a connector within the sensorimotor-auditory network typically recruited during speech processing. However, when comprehension is compromised due to degradation of speech input, the area disengages from the sensorimotor-auditory network. It becomes part of the visual network again and turns from connector into a simple peripheral node. These varying connectivity patterns are coherent with the Interactive Account considering the left-vOT as a convergent zone with multiple functions and interaction patterns that depend on task demands and the nature of sensory input.

neuroscience↗

bioRxiv2020

The cerebellum is involved in internal and external speech error monitoring

An fMRI study examined how speakers inspect their own speech for errors. In a word production task, we observed enhanced involvement of the right posterior cerebellum for trials that were correct, but on which participants were more likely to make a word-as compared to a non-word error. Furthermore, comparing errors to correctly produced utterances, we observed increased activation of the same cerebellar region, in addition to temporal and medial frontal regions. Within the framework associating the cerebellum to forward modelling of upcoming actions, this indicates that forward models of verbal actions contain information about word representations used for error monitoring even before articulation (internal monitoring). Additional resources relying on speech perception and conflict monitoring are deployed during articulation to detect overt errors (external monitoring). In summary, speech monitoring seems to recruit a network of brain regions serving domain general purposes, even for abstract levels of processing.

neuroscience↗