This webinar convened researchers at the intersection of Artificial Intelligence and Neuroscience to investigate how large language models (LLMs) can serve as valuable “model organisms” for understanding human language processing. Presenters showcased evidence that brain recordings (fMRI, MEG, ECoG) acquired while participants read or listened to unconstrained speech can be predicted by representations extracted from state-of-the-art text- and speech-based LLMs. In particular, text-based LLMs tend to align better with higher-level language regions, capturing more semantic aspects, while speech-based LLMs excel at explaining early auditory cortical responses. However, purely low-level features can drive part of these alignments, complicating interpretations. New methods, including perturbation analyses, highlight which linguistic variables matter for each cortical area and time scale. Further, “brain tuning” of LLMs—fine-tuning on measured neural signals—can improve semantic representations and downstream language tasks. Despite open questions about interpretability and exact neural mechanisms, these results demonstrate that LLMs provide a promising framework for probing the computations underlying human language comprehension and production at multiple spatiotemporal scales.

The ability to build, store, and manipulate semantic representations lies at the core of all our (inter)actions. Combining evidence from cognitive neuroimaging and experimental neuropsychology, I study the neurocognitive correlates of semantic knowledge in relation to other cognitive functions, chiefly language. In this talk, I will start by reviewing neuroimaging findings supporting the idea that semantic representations are encoded in distributed yet specialized cortical areas (1), and rapidly recovered (2) according to the requirement of the task at hand (3). I will then focus on studies conducted in neurodegenerative patients, offering a unique window on the key role played by a structurally and functionally heterogeneous piece of cortex: the anterior temporal lobe (4,5). I will present pathological, neuroimaging, cognitive, and behavioral data illustrating how damages to language-related networks can affect or spare semantic knowledge as well as possible paths to functional compensation (6,7). Time permitting, we will discuss the neurocognitive dissociation between nouns and verbs (8) and how verb production is differentially impacted by specific language impairments (9).

Hongjing Lu will present a new computational model of Probabilistic Analogical Mapping (PAM, in collaboration with Nick Ichien and Keith Holyoak) that finds systematic correspondences between inputs generated by machine learning. The model adopts a Bayesian framework for probabilistic graph matching, operating on semantic relation networks constructed from distributed representations of individual concepts (word embeddings created by Word2vec) and of relations between concepts (created by our BART model). We have used PAM to simulate a broad range of phenomena involving analogical mapping by both adults and children. Our approach demonstrates that human-like analogical mapping can emerge from comparison mechanisms applied to rich semantic representations of individual concepts and relations. More details can be found https://arxiv.org/ftp/arxiv/papers/2103/2103.16704.pdf

The ventral visual pathway extends from the occipital to the anterior temporal regions, and is specialized in giving meaning to objects and people that are perceived through vision. Numerous studies in functional magnetic resonance imaging have focused on the cerebral basis of visual recognition. However, this technique is susceptible to magnetic artefacts in ventral anterior temporal regions and it has led to an underestimation of the role of these regions within the ventral visual stream, especially with respect to face recognition and semantic representations. Moreover, there is an increasing need for implicit methods assessing these functions as explicit tasks lack specificity. In this talk, I will present three studies using fast periodic visual stimulation (FPVS) in combination with scalp and/or intracerebral EEG to overcome these limitations and provide high SNR in temporal regions. I will show that, beyond face recognition, FPVS can be extended to investigate semantic representations using a face-name association paradigm and a semantic categorisation paradigm with written words. These results shed new light on the role of temporal regions and demonstrate the high potential of the FPVS approach as a powerful electrophysiological tool to assess various cognitive functions in neurotypical and clinical populations.