Brain decoding: Opportunities and challenges for pattern recognition

Dimitri Van De Ville, Seong Whan Lee

Research output: Contribution to journalEditorial

9 Citations (Scopus)

Abstract

The neuroimaging community heavily relies on statistical inference to explain measured brain activity given the experimental paradigm. Undeniably, this method has led to many results, but it is limited by the richness of the generative models that are deployed, typically in a mass-univariate way. Such an approach is suboptimal given the high-dimensional and complex spatiotemporal correlation structure of neuroimaging data. Over the recent years, techniques from pattern recognition have brought new insights into where and how information is stored in the brain by prediction of the stimulus or state from the data. Pattern recognition is intrinsically multivariate and the underlying models are data-driven. Moreover, the predictive setting is more powerful for many applications, including clinical diagnosis and braincomputer interfacing. This special issue features a number of papers that identify and tackle remaining challenges in this field. The specific problems at hand constitute opportunities for future research in pattern recognition and neurosciences.

Original languageEnglish
Pages (from-to)2033-2034
Number of pages2
JournalPattern Recognition
Volume45
Issue number6
DOIs
Publication statusPublished - 2012 Jun

    Fingerprint

Keywords

  • Brain decoding
  • Braincomputer interface
  • Electroencephalography
  • Functional magnetic resonance imaging
  • Neuroimaging

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this