skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Spectral Transfer Learning Using Information Geometry for a User-Independent Brain-Computer Interface

Abstract

Recent advances in signal processing and machine learning techniques have enabled the application of Brain-Computer Interface (BCI) technologies to fields such as medicine, industry, and recreation; however, BCIs still suffer from the requirement of frequent calibration sessions due to the intra- and inter-individual variability of brain-signals, which makes calibration suppression through transfer learning an area of increasing interest for the development of practical BCI systems. In this paper, we present an unsupervised transfer method (spectral transfer using information geometry,STIG),which ranks and combines unlabeled predictions from an ensemble of information geometry classifiers built on data from individual training subjects. The STIG method is validated in both off-line and real-time feedback analysis during a rapid serial visual presentation task (RSVP). For detection of single-trial, event-related potentials (ERPs), the proposed method can significantly outperform existing calibration-free techniques as well as out perform traditional within-subject calibration techniques when limited data is available. Here, this method demonstrates that unsupervised transfer learning for single-trial detection in ERP-based BCIs can be achieved without the requirement of costly training data, representing a step-forward in the overall goal of achieving a practical user-independent BCI system.

Authors:
 [1];  [2];  [3];  [4];  [4]
  1. U.S. Army Research Lab., Aberdeen Proving Ground, MD (United States); Columbia Univ., New York, NY (United States)
  2. U.S. Army Research Lab., Aberdeen Proving Ground, MD (United States); Univ. of Texas, San Antonio, TX (United States)
  3. U.S. Army Research Lab., Aberdeen Proving Ground, MD (United States); Univ. of Maryland, College Park, MD (United States)
  4. U.S. Army Research Lab., Aberdeen Proving Ground, MD (United States)
Publication Date:
Research Org.:
Columbia Univ., New York, NY (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1378936
Resource Type:
Journal Article: Accepted Manuscript
Journal Name:
Frontiers in Neuroscience (Online)
Additional Journal Information:
Journal Name: Frontiers in Neuroscience (Online); Journal Volume: 10; Journal ID: ISSN 1662-453X
Publisher:
Frontiers Research Foundation
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING; 60 APPLIED LIFE SCIENCES; unsupervised learning; ensemble learning; calibration-free BCI; P300; RSVP

Citation Formats

Waytowich, Nicholas R., Lawhern, Vernon J., Bohannon, Addison W., Ball, Kenneth R., and Lance, Brent J. Spectral Transfer Learning Using Information Geometry for a User-Independent Brain-Computer Interface. United States: N. p., 2016. Web. doi:10.3389/fnins.2016.00430.
Waytowich, Nicholas R., Lawhern, Vernon J., Bohannon, Addison W., Ball, Kenneth R., & Lance, Brent J. Spectral Transfer Learning Using Information Geometry for a User-Independent Brain-Computer Interface. United States. doi:10.3389/fnins.2016.00430.
Waytowich, Nicholas R., Lawhern, Vernon J., Bohannon, Addison W., Ball, Kenneth R., and Lance, Brent J. 2016. "Spectral Transfer Learning Using Information Geometry for a User-Independent Brain-Computer Interface". United States. doi:10.3389/fnins.2016.00430. https://www.osti.gov/servlets/purl/1378936.
@article{osti_1378936,
title = {Spectral Transfer Learning Using Information Geometry for a User-Independent Brain-Computer Interface},
author = {Waytowich, Nicholas R. and Lawhern, Vernon J. and Bohannon, Addison W. and Ball, Kenneth R. and Lance, Brent J.},
abstractNote = {Recent advances in signal processing and machine learning techniques have enabled the application of Brain-Computer Interface (BCI) technologies to fields such as medicine, industry, and recreation; however, BCIs still suffer from the requirement of frequent calibration sessions due to the intra- and inter-individual variability of brain-signals, which makes calibration suppression through transfer learning an area of increasing interest for the development of practical BCI systems. In this paper, we present an unsupervised transfer method (spectral transfer using information geometry,STIG),which ranks and combines unlabeled predictions from an ensemble of information geometry classifiers built on data from individual training subjects. The STIG method is validated in both off-line and real-time feedback analysis during a rapid serial visual presentation task (RSVP). For detection of single-trial, event-related potentials (ERPs), the proposed method can significantly outperform existing calibration-free techniques as well as out perform traditional within-subject calibration techniques when limited data is available. Here, this method demonstrates that unsupervised transfer learning for single-trial detection in ERP-based BCIs can be achieved without the requirement of costly training data, representing a step-forward in the overall goal of achieving a practical user-independent BCI system.},
doi = {10.3389/fnins.2016.00430},
journal = {Frontiers in Neuroscience (Online)},
number = ,
volume = 10,
place = {United States},
year = 2016,
month = 9
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Save / Share:
  • This article describes ten of the most frequent ergonomic problems found in human--computer interfaces (HCIs) associated with complex industrial machines. In contrast with being thought of a ``user friendly,`` many of these machines are seen as exhibiting ``user-hostile`` attributes by the author. The historical lack of consistent application of ergonomic principles in the HCIs has led to a breed of very sophisticated, complex manufacturing systems that few people can operate without extensive orientation, training, or experience. This design oversight has produced the need for customized training programs and help documentation, unnecessary machine downtime, and reduced productivity resulting from operator stressmore » and confusion. The ten issues are treated in a problem--solution format with real-world graphic examples of good and poor design. Intended for a diverse audience, the article avoids technical jargon, and is appropriate reading for those involved in software, product engineering, marketing, and management. {copyright} {ital 1995} {ital American} {ital Vacuum} {ital Society}« less
  • A graphical user interface, using the Interactive Data Language (IDL) widget toolkit, for calculation of spectral properties of synchrotron radiation sources and for interaction of x-rays with optical elements has been developed. The interface runs presently on three different computer architectures under the Unix operating system {endash} the Sun-OS, the HP-UX, and the DEC-Unix operating systems. The point-and-click interface is used as a driver program for a variety of codes from different authors written in different computer languages. The execution of codes for calculating synchrotron radiation from undulators, wigglers, and bending magnets is summarized. The computation of optical properties ofmore » materials and the x-ray diffraction profiles from crystals in different geometries are also discussed. The interface largely simplifies the use of these codes and may be used without prior knowledge of how to run a particular program. {copyright} {ital 1996 American Institute of Physics.}« less
  • No abstract prepared.
  • In this paper, we present the discovery and preliminary characterization of a gravitationally lensed quasar with a source redshift z s = 2.74 and image separation of 2.9 arcsec lensed by a foreground z l = 0.40 elliptical galaxy. Since optical observations of gravitationally lensed quasars show the lens system as a superposition of multiple point sources and a foreground lensing galaxy, we have developed a morphology-independent multi-wavelength approach to the photometric selection of lensed quasar candidates based on Gaussian Mixture Models (GMM) supervised machine learning. Using this technique and gi multicolour photometric observations from the Dark Energy Survey (DES),more » near-IR JK photometry from the VISTA Hemisphere Survey (VHS) and WISE mid-IR photometry, we have identified a candidate system with two catalogue components with i AB = 18.61 and i AB = 20.44 comprising an elliptical galaxy and two blue point sources. Spectroscopic follow-up with NTT and the use of an archival AAT spectrum show that the point sources can be identified as a lensed quasar with an emission line redshift of z = 2.739 ± 0.003 and a foreground early-type galaxy with z = 0.400 ± 0.002. We model the system as a single isothermal ellipsoid and find the Einstein radius θ E ~ 1.47 arcsec, enclosed mass M enc ~ 4 × 10 11 M and a time delay of ~52 d. Finally, the relatively wide separation, month scale time delay duration and high redshift make this an ideal system for constraining the expansion rate beyond a redshift of 1.« less