The Brain's Bitter Lesson: Scaling Speech Decoding With Self-Supervised Learning
- URL: http://arxiv.org/abs/2406.04328v4
- Date: Fri, 31 Jan 2025 23:00:26 GMT
- Title: The Brain's Bitter Lesson: Scaling Speech Decoding With Self-Supervised Learning
- Authors: Dulhan Jayalath, Gilad Landau, Brendan Shillingford, Mark Woolrich, Oiwi Parker Jones,
- Abstract summary: We develop self-supervised objectives, together with an architecture, for learning from heterogeneous brain recordings.
Scaling to nearly 400 hours of MEG data and 900 subjects, our approach shows generalisation across participants, datasets, tasks, and even to novel subjects.
It achieves improvements of 15-27% over state-of-the-art models and matches surgical decoding performance with non-invasive data.
- Score: 3.649801602551928
- License:
- Abstract: The past few years have seen remarkable progress in the decoding of speech from brain activity, primarily driven by large single-subject datasets. However, due to individual variation, such as anatomy, and differences in task design and scanning hardware, leveraging data across subjects and datasets remains challenging. In turn, the field has not benefited from the growing number of open neural data repositories to exploit large-scale deep learning. To address this, we develop neuroscience-informed self-supervised objectives, together with an architecture, for learning from heterogeneous brain recordings. Scaling to nearly 400 hours of MEG data and 900 subjects, our approach shows generalisation across participants, datasets, tasks, and even to novel subjects. It achieves improvements of 15-27% over state-of-the-art models and matches surgical decoding performance with non-invasive data. These advances unlock the potential for scaling speech decoding models beyond the current frontier.
Related papers
- Resolving Domain Shift For Representations Of Speech In Non-Invasive Brain Recordings [3.5297361401370044]
We focus on non-invasive data collected using magnetoencephalography (MEG)
To the best of our knowledge, this study is the first ever application of feature-level, deep learning based on MEG neuroimaging data.
arXiv Detail & Related papers (2024-10-25T21:56:23Z) - Knowledge-Guided Prompt Learning for Lifespan Brain MR Image Segmentation [53.70131202548981]
We present a two-step segmentation framework employing Knowledge-Guided Prompt Learning (KGPL) for brain MRI.
Specifically, we first pre-train segmentation models on large-scale datasets with sub-optimal labels.
The introduction of knowledge-wise prompts captures semantic relationships between anatomical variability and biological processes.
arXiv Detail & Related papers (2024-07-31T04:32:43Z) - BrainSegFounder: Towards 3D Foundation Models for Neuroimage Segmentation [6.5388528484686885]
This study introduces a novel approach towards the creation of medical foundation models.
Our method involves a novel two-stage pretraining approach using vision transformers.
BrainFounder demonstrates a significant performance gain, surpassing the achievements of previous winning solutions.
arXiv Detail & Related papers (2024-06-14T19:49:45Z) - Aligning brain functions boosts the decoding of visual semantics in
novel subjects [3.226564454654026]
We propose to boost brain decoding by aligning brain responses to videos and static images across subjects.
Our method improves out-of-subject decoding performance by up to 75%.
It also outperforms classical single-subject approaches when fewer than 100 minutes of data is available for the tested subject.
arXiv Detail & Related papers (2023-12-11T15:55:20Z) - Deep Learning for real-time neural decoding of grasp [0.0]
We present a Deep Learning-based approach to the decoding of neural signals for grasp type classification.
The main goal of the presented approach is to improve over state-of-the-art decoding accuracy without relying on any prior neuroscience knowledge.
arXiv Detail & Related papers (2023-11-02T08:26:29Z) - Fighting the scanner effect in brain MRI segmentation with a progressive
level-of-detail network trained on multi-site data [1.6379393441314491]
LOD-Brain is a 3D convolutional neural network with progressive levels-of-detail able to segment brain data from any site.
It produces state-of-the-art results, with no significant difference in performance between internal and external sites.
Its portability opens the way for large scale application across different healthcare institutions, patient populations, and imaging technology manufacturers.
arXiv Detail & Related papers (2022-11-04T12:15:18Z) - Decoding speech perception from non-invasive brain recordings [48.46819575538446]
We introduce a model trained with contrastive-learning to decode self-supervised representations of perceived speech from non-invasive recordings.
Our model can identify, from 3 seconds of MEG signals, the corresponding speech segment with up to 41% accuracy out of more than 1,000 distinct possibilities.
arXiv Detail & Related papers (2022-08-25T10:01:43Z) - 2021 BEETL Competition: Advancing Transfer Learning for Subject
Independence & Heterogenous EEG Data Sets [89.84774119537087]
We design two transfer learning challenges around diagnostics and Brain-Computer-Interfacing (BCI)
Task 1 is centred on medical diagnostics, addressing automatic sleep stage annotation across subjects.
Task 2 is centred on Brain-Computer Interfacing (BCI), addressing motor imagery decoding across both subjects and data sets.
arXiv Detail & Related papers (2022-02-14T12:12:20Z) - Overcoming the Domain Gap in Neural Action Representations [60.47807856873544]
3D pose data can now be reliably extracted from multi-view video sequences without manual intervention.
We propose to use it to guide the encoding of neural action representations together with a set of neural and behavioral augmentations.
To reduce the domain gap, during training, we swap neural and behavioral data across animals that seem to be performing similar actions.
arXiv Detail & Related papers (2021-12-02T12:45:46Z) - Deep Recurrent Encoder: A scalable end-to-end network to model brain
signals [122.1055193683784]
We propose an end-to-end deep learning architecture trained to predict the brain responses of multiple subjects at once.
We successfully test this approach on a large cohort of magnetoencephalography (MEG) recordings acquired during a one-hour reading task.
arXiv Detail & Related papers (2021-03-03T11:39:17Z) - Neural Data-to-Text Generation via Jointly Learning the Segmentation and
Correspondence [48.765579605145454]
We propose to explicitly segment target text into fragment units and align them with their data correspondences.
The resulting architecture maintains the same expressive power as neural attention models.
On both E2E and WebNLG benchmarks, we show the proposed model consistently outperforms its neural attention counterparts.
arXiv Detail & Related papers (2020-05-03T14:28:28Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.