NEURAL MEASURES FOR HCI EVALUATION

Alan Gevins

SAM Technology and EEG Systems Lab
101 Spear St. #203
San Francisco, CA 94105

CONTACT INFORMATION

Alan Gevins, SAM Technology and EEG Systems Lab, 101 Spear St. #203, San Francisco, CA 94105

alan@eeg.com
415-227-4900 X133
FAX 415-546-7122

PROGRAM AREA

Usability and User-Centered Design.

KEYWORDS

HCI, usability evaluation, brain activity, EEG

PROJECT SUMMARY

A serious obstacle in developing better human-computer interfaces is the lack of a convenient objective means for assessing whether one type of interface requires more mental effort than another. In this regard, it would seem that neurophysiological measures such as the electroencephalogram (EEG) could be used to derive accurate and unobtrusive assessments of the degree of a user+s mental effort as they worked at a computer. There are, however, formidable scientific and technical barriers that have heretofore prevented the systematic use of EEG indices in HCI research. We have been working to overcome all aspects of these impediments. Specifically, in this project, we have been developing algorithms to measure neural signals of mental effort which are not confounded by a variety of artifactual contaminants. We have been evaluating the sensitivity and specificity of these measures with empirical studies of subjects performing highly controlled computer-based tasks. The results so far strongly suggest that accurate and reliable inferences about the degree of an individual+s mental effort can be derived from neurophysiological measures in a practical fashion.

In one recent study, features of the EEG related to mental effort were identified and extracted and effective automated means for classifying the features according to degree of task difficulty were developed and tested. To do this, an experiment was designed to systematically vary task difficulty (in terms of increasing working memory load) while holding stimulus- and response-related factors constant. Performance and electrophysiological data were collected from eight subjects while they performed tasks that taxed spatial and verbal working memory, functions which are critical for performance of most complex computer-based tasks. When neural-network pattern recognition classification functions were applied to automatically discriminate high and low workload levels, over 95% accuracy was obtained on independent testing data. Further, networks trained on data from one day were found to reliably classify data collected on another day, networks trained on data from one task were found to be general enough to accurately classify data collected in another similarly well-controlled task, and networks trained on data from a group of subjects were found to be general enough to classify data from new subjects who were not part of the training group.

Supported by NSF and AFOSR

PROJECT REFERENCES

Gevins, A.S., Smith, M.E., Le, J., Leong, H., Bennett, J., Martin, N., McEvoy, L., Du., R., and Whitfield, S. (1995) High resolution evoked potential imaging of the cortical dynamics of human working memory. Electroencephalography and Clinical Neurophysiology, In Press.

Gevins, A., Leong, H., Du, R., Smith, M.E. et al. (1995). Towards measurement of brain function in operational environments. Biological Psychology, 40, 169-186.

Gevins, A., Leong, H., and Smith, M.E., Le, J. and Du, R. (1995). Mapping cognitive brain function with modern high resolution electroencephalography. Trends in Neurosciences, 18, 429-436.

Gevins, A.S. and Cutillo, B.A. (1993) Spatiotemporal dynamics of component processes in human working memory. Electroencephalography and Clinical Neurophysiology, 87, 128-143.