Transfer Learning in Brain-Computer Interfaces

TL;DR

This paper reviews transfer learning techniques in BCIs, introduces a flexible framework and a novel EEG-specific regression method, demonstrating improved transfer performance across subjects and sessions.

Contribution

It presents a general transfer learning framework for BCIs and a new EEG-structured regression method that outperforms existing approaches.

Findings

Framework improves subject-to-subject transfer in motor imagery

Method outperforms comparable techniques on ALS patient data

Effective transfer across sessions and subjects demonstrated

Abstract

The performance of brain-computer interfaces (BCIs) improves with the amount of available training data, the statistical distribution of this data, however, varies across subjects as well as across sessions within individual subjects, limiting the transferability of training data or trained models between them. In this article, we review current transfer learning techniques in BCIs that exploit shared structure between training data of multiple subjects and/or sessions to increase performance. We then present a framework for transfer learning in the context of BCIs that can be applied to any arbitrary feature space, as well as a novel regression estimation method that is specifically designed for the structure of a system based on the electroencephalogram (EEG). We demonstrate the utility of our framework and method on subject-to-subject transfer in a motor-imagery paradigm as well as on session-to-session transfer in one patient diagnosed with amyotrophic lateral sclerosis (ALS), showing that it is able to outperform other comparable methods on an identical dataset.