Summary of Fmri Predictors Based on Language Models Of Increasing Complexity Recover Brain Left Lateralization, by Laurent Bonnasse-gahot and Christophe Pallier
fMRI predictors based on language models of increasing complexity recover brain left lateralization
by Laurent Bonnasse-Gahot, Christophe Pallier
First submitted to arxiv on: 28 May 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Neurons and Cognition (q-bio.NC)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed study analyzes naturalistic language processing by scanning participants while listening to continuous text, employing encoding models that identify significant correlations between brain signals and model predictions. The research reveals symmetric bilateral activation patterns, contradicting traditional left lateralization of language processing. This paper presents a novel analysis of fMRI datasets, testing 28 large language models with varying complexity (124M-14.2B parameters). The findings demonstrate a scaling law relationship between model performance in predicting brain responses and the logarithm of its parameter count. Interestingly, this effect is stronger in the left hemisphere, reconciling computational models with classic aphasic patient studies. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A team of researchers used special machines to scan people’s brains while they listened to continuous text. They wanted to understand how our brains process language naturally. The study showed that some areas of the brain work together on both sides, which is different from what we thought about language processing. This paper explores a big dataset with 28 models that can process language in different ways. They found that the better the model is at predicting brain activity, the more complex it needs to be. The study also shows that the left side of the brain plays a bigger role than the right side in this process. |
