Language and music are two human-unique capacities whose relationship remains debated. Some have argued for overlap in processing mechanisms, especially structure processing. Such claims often concern the inferior frontal component of language system located within "Broca's area." However, others failed to find overlap. Using a robust individual-subject fMRI approach, we examined responses brain regions stimuli, probed musical abilities individuals with severe aphasia. Across 4 experiments,...

Abstract A network of left frontal and temporal brain regions supports language processing. This “core” stores our knowledge words constructions as well constraints on how those combine to form sentences. However, linguistic additionally includes information about phonemes they phonemic clusters, syllables, words. Are phoneme combinatorics also represented in these regions? Across five functional magnetic resonance imaging experiments, we investigated the sensitivity high-level processing...

Focal adhesions (FAs) are specialized membrane-associated multi-protein complexes that link the cell to extracellular matrix and play crucial roles in cell-matrix sensing. Considerable information is available on complex molecular composition of these sites, yet regulation FA dynamics largely unknown. Based a combination FRAP studies live cells, with silico simulations mathematical modeling, we show plaque proteins paxillin vinculin exist four dynamic states: an immobile FA-bound fraction,...

Abstract Language and music are two human-unique capacities whose relationship remains debated. Some have argued for overlap in processing mechanisms, especially structure processing. Such claims often concern the inferior frontal component of language system located within ‘Broca’s area’. However, others failed to find overlap. Using a robust individual-subject fMRI approach, we examined responses brain regions stimuli, probed musical abilities individuals with severe aphasia. Across four...

Conversation is a dynamic, multi-modal activity involving the exchange of complex streams information like words, prosody, gesture, eye contact, and backchannels. Understanding how these different channels interact in naturalistic scenarios essential for understanding mechanisms governing human communication.Past studies suggested that duration words tied to their predictability context, but it remains unclear whether this relationship speaker-oriented (e.g. retrieval or production-based)...

Conversation is a dynamic, multi-modal activity involving the exchange of complex streams information like words, prosody, gesture, eye contact, and backchannels. Understanding how these different channels interact in naturalistic scenarios essential for understanding mechanisms governing human communication.Past studies suggested that duration words tied to their predictability context, but it remains unclear whether this relationship speaker-oriented (e.g. retrieval or production-based)...

Much of what is known about the timing visual processing in brain inferred from intracranial studies monkeys, with human data limited to mainly noninvasive methods lower spatial resolution. Here, we estimated onset latencies electrocorticographic (ECoG) recordings a patient who was implanted 112 subdural electrodes, distributed across posterior cortex right hemisphere, for presurgical evaluation intractable epilepsy. Functional MRI prior surgery used determine boundaries areas. The presented...

Background DNA chips allow simultaneous measurements of genome-wide response thousands genes, i.e. system level monitoring the gene-network activity. Advanced analysis methods have been developed to extract meaningful information from vast amount raw gene-expression data obtained microarray measurements. These usually aimed distinguish between groups subjects (e.g., cancer patients vs. healthy subjects) or identifying marker genes that help those groups. We assumed motifs related internal...

Abstract Despite long knowing what brain areas support language comprehension, our knowledge of the neural computations that these frontal and temporal regions implement remains limited. One important unresolved question concerns functional differences among populations comprise network. Leveraging high spatiotemporal resolution intracranial recordings, we examined responses to sentences linguistically degraded conditions discovered three response profiles differ in their dynamics. These...

ABSTRACT A network of left frontal and temporal brain regions supports ‘high-level’ language processing— including the processing word meanings, as well word-combinatorial processing—across presentation modalities. This ‘core’ has been argued to store our knowledge words constructions constraints on how those combine form sentences. However, linguistic additionally includes information about sounds (phonemes) they clusters, syllables, words. Is this phoneme combinatorics also represented in...

The perceptual organization of pitch is frequently described as helical, with a monotonic dimension height and circular chroma, accounting for the repeating structure octave. Although neural representation widely studied, way in which chroma manifested activity currently debated. We tested automaticity processing using MMN-an ERP component indexing automatic detection deviations from auditory regularity. Musicians trained to classify pure or complex tones across four octaves, based on...

Abstract Everyday auditory streams are complex, including spectro-temporal content that varies at multiple timescales. Using EEG, we investigated the sensitivity of human cortex to past stimulation in unattended sequences equiprobable tones. In 3 experiments 82 participants overall, found neural responses measured different latencies after stimulus onset were sensitive frequency intervals computed over distinct Importantly, early a longer history than later responses. To account for these...

Prosody—the suprasegmental component of speech, including pitch, loudness, and tempo—carries critical aspects meaning. However, the relationship between information conveyed by prosody vs. words themselves remains poorly understood. We use large language models (LLMs) to estimate how much is redundant themselves. Using a spoken corpus English audiobooks, we extract prosodic features aligned individual test well they can be predicted from LLM embeddings, compared non-contextual word...

Lukas Wolf, Klemen Kotar, Greta Tuckute, Eghbal Hosseini, Tamar I. Regev, Ethan Gotlieb Wilcox, Alexander Scott Warstadt. Proceedings of the BabyLM Challenge at 27th Conference on Computational Natural Language Learning. 2023.

The Supporting Information text file does not appear. It should be cited as Text S1, with a caption Mathematical modeling and derivation of analytical expression, supplementary figures legends. Please view S1 here: Click here for additional data file.(70K, pdf)

ABSTRACT Everyday auditory streams are complex, including spectro-temporal content that varies at multiple time scales. Using EEG, we investigate the sensitivity of human cortex to past stimulation in unattended sequences equiprobable tones. In 3 experiments 82 participants overall, found neural responses measured different latencies after stimulus onset were sensitive frequency intervals computed over distinct Importantly, early a longer history than later responses. To account for these...

Abstract Much of what is known about the timing visual processing in brain inferred from intracranial studies monkeys, with human data limited to mainly non-invasive methods lower spatial resolution. Here, we estimated onset latencies electrocorticographic (ECoG) recordings a patient who was implanted 112 sub-dural electrodes, distributed across posterior cortex right hemisphere, for pre-surgical evaluation intractable epilepsy. Functional MRI prior surgery used determine boundaries areas....

Human cortical responses to natural sounds, measured with fMRI, can be approximated as the weighted sum of a small number canonical response patterns (components), each having interpretable functional and anatomical properties. Here, we asked whether this organization is preserved in cases where only one temporal lobe available due early brain damage by investigating unique family: sibling born without left lobe, another right third anatomically neurotypical. We analyzed fMRI diverse sounds...

Training on multiple modalities of input can augment the capabilities a language model. Here, we ask whether such training regime improve quality and efficiency these systems as well. We focus text--audio introduce Whisbert, which is inspired by text--image approach FLAVA (Singh et al., 2022). In accordance with Babylm guidelines (Warstadt 2023), pretrain Whisbert dataset comprising only 100 million words plus their corresponding speech from word-aligned version People's Speech (Galvez...