Abstraction Induces the Brain Alignment of Language and Speech Models

Emily Cheng, Aditya R. Vaidya, Richard Antonello

TL;DR

Why do the intermediate layers of large language and speech models predict human brain activity better than their final output layers? This paper demonstrates that model-brain alignment is driven by meaning abstraction—quantified by a peak in the representation's intrinsic dimension—rather than the next-word prediction objective itself.

What Problem Does This Paper Address?

A fascinating discovery in computational neuroscience is that the hidden states of deep language models (LLMs) and speech models can linearly predict neural responses (measured via fMRI or ECoG) when humans listen to natural language. However, an unexplained anomaly persists: the intermediate layers of these networks consistently outperform the final output layers at predicting brain activity.

This raises a fundamental question about why models and brains align. Is it because they share a similar predictive coding objective (minimizing next-word surprisal), or is it a byproduct of convergent evolution toward shared abstractions of meaning? This paper addresses this gap by investigating the geometric and semantic properties of hidden layers across various language and audio foundation models.

Methods

The authors employ a multi-pronged approach combining neural encoding, information theory, and representation geometry:

  1. Brain Encoding Models: They train ridge regression models to map hidden states from LLMs and speech models to fMRI and ECoG brain signals.
  2. Intrinsic Dimension ($I_d$): To measure feature complexity without assuming linearity, they use a state-of-the-art nonlinear estimator called GRIDE (Generalized Ratios Intrinsic Dimension Estimator) on generic, in-distribution datasets.
  3. Layerwise Probing: They use linear probes (SentEval for text, GloVe/Mel-filterbanks for audio) to evaluate whether a layer contains low-level surface features or high-level semantic information.
  4. Layerwise Surprisal: They use the TunedLens technique to map intermediate layers directly to the vocabulary space, calculating next-token prediction error at each stage of processing.
  5. Causal Manipulation (Braintuning): They finetune a speech model directly on fMRI data to observe how optimizing for brain predictivity alters the model's internal geometry and semantic content.

Datasets

The study utilizes several open-source datasets across modalities:

  • fMRI: Data from 3 subjects listening to ~20 hours of English podcast stories (LeBel et al., 2023).
  • ECoG: High-gamma frequency band signals from 9 subjects listening to a 30-minute podcast (Zada et al., 2025).
  • Model Training/Evaluation: The Pile (for LLMs), LibriSpeech (for speech models), and SentEval (for linguistic probing tasks).

Key Insights and Ideas

  • The Two-Phase Processing Hypothesis: Language models operate in two distinct phases across their layers. First, they build up complex, higher-order linguistic features from the input, causing the intrinsic dimension ($I_d$) to peak in the middle layers. Second, they compress and refine this representation to fit the specific output distribution required for next-token prediction.
  • Prediction Narrows Representations: While final layers are optimal for minimizing next-token surprisal, this objective actually forces a dimensionality bottleneck that strips away broader semantic abstractions, making output layers less similar to the human brain.
  • Abstraction, Not Just Dimensionality: High dimensionality alone does not guarantee brain alignment. Testing a baseline of Random Fourier Features (RFF) showed that while increasing dimensionality improves brain predictivity slightly, it plateaus at a very low level compared to actual language models. The dimensionality must represent structured linguistic abstraction.
Figure 3. Semantic decodability coalesces near the Id peak. For LLMs (top rows) and speech models (bottom), higher-order feature decodability (purple) peaks in the deep layers near the Id peak (dashed line); superficial text or acoustic feature decodability (pink) drops over the layers. Task details in Appendix D.

Key Results

  • $I_d$ Strongly Tracks Brain Predictivity: Across virtually all tested models (OPT, Pythia, WavLM, Whisper), the layers with the highest intrinsic dimension were almost perfectly co-located with the layers that best predicted fMRI and ECoG signals. The global Spearman correlation between normalized $I_d$ and encoding performance was $\rho = 0.76$ for fMRI and $\rho = 0.43$ for ECoG.
  • Surprisal Fails to Explain Layerwise Alignment: Next-token surprisal decreases monotonically across layers, meaning the final layer is always the best predictor of the next word. Yet, brain predictivity peaks much earlier (between 40-70% of processing). In mixed-effects modeling, $I_d$ completely explained away the effect of surprisal for fMRI data.
Figure 1. LLM surprisal and encoding performance. (A) Surprisal over LLM layers. For each LLM, the next-token surprisal from a given layer (y-axis), fit with an affine map (Belrose et al., 2023), decreases monotonically over processing (x-axis). The lowest surprisal layer is always the last one, while the layer most predictive of the brain (◦) occurs between 40-70% of processing. (B) Layerwise encoding performance vs. layerwise surprisal. For each LLM, processing trajectories through the (encoding performance, surprisal) plane are shown, where surprisals are normalized by the log-vocabulary size for comparability between models. The global Spearman correlation across models is ρ = −0.63 for fMRI and ρ = −0.31 for ECoG, significant at a p-value cutoff of 0.01 (*).
  • Causal Evidence via Braintuning: Directly finetuning WavLM to better predict fMRI responses causally increased both its internal intrinsic dimension and its performance on semantic probing tasks, confirming a tight, non-trivial link between semantic richness, geometry, and brain alignment.

Limitations and Future Work

The authors note that Intrinsic Dimension ($I_d$) is a coarse-grained, dataset-level metric. While it successfully quantifies the number of degrees of freedom (or feature complexity) within a layer, it does not explicitly reveal what those features are (e.g., specific syntactic structures vs. semantic concepts).

Future work is needed to explicitly map these geometric degrees of freedom to interpretable linguistic properties and to fully untangle the co-dependencies between intrinsic dimensionality and neural encoding performance.

Summary updated

Abstract

Research has repeatedly demonstrated that intermediate hidden states extracted from large language models and speech audio models predict measured brain response to natural language stimuli. Yet, very little is known about the representation properties that enable this high prediction performance. Why is it the intermediate layers, and not the output layers, that are most effective for this unique and highly general transfer task? We give evidence that the correspondence between speech and language models and the brain derives from shared meaning abstraction and not their next-word prediction properties. In particular, models construct higher-order linguistic features in their middle layers, cued by a peak in the layerwise intrinsic dimension, a measure of feature complexity. We show that a layer's intrinsic dimension strongly predicts how well it explains fMRI and ECoG signals; that the relation between intrinsic dimension and brain predictivity arises over model pre-training; and finetuning models to better predict the brain causally increases both representations' intrinsic dimension and their semantic content. Results suggest that semantic richness, high intrinsic dimension, and brain predictivity mirror each other, and that the key driver of model-brain similarity is rich meaning abstraction of the inputs, where language modeling is a task complex enough (but perhaps not the only) to require it.
BibTeX
@inproceedings{cheng2026abstraction,
  author = {Emily Cheng and Aditya R. Vaidya and Richard Antonello},
  title = {{Abstraction Induces the Brain Alignment of Language and Speech Models}},
  booktitle = {International Conference on Machine Learning},
  year = {2026},
  url = {https://openreview.net/forum?id=n5Ds4qbtjM},
}
Figure