Relationship between working memory performance and neural activation measured using near-infrared spectroscopy

Working memory (WM) is a key function for various cognitive processes. Near-infrared spectroscopy (NIRS) is a powerful technique for noninvasive functional imaging. However, a study has yet to be published on the application of NIRS for evaluating WM performance. The objective was to evaluate NIRS for measuring WM performance.

Subjects were trained to perform a visuospatial WM task. Eight channels on the lateral prefrontal cortex were analyzed. We asked the following three questions: (1) Does WM performance correlate with NIRS signal amplitudes? (2) What are the differences in NIRS amplitudes between correct- and incorrect-WM tasks? (3) Is there a correlation between WM performance and NIRS amplitudes in only correct-WM tasks?

NIRS activation in all channels correlated with WM performance (P < 0.05). There was a statistically significant difference (P < 0.05) in seven channels between NIRS amplitude in correct- and incorrect-WM tasks. NIRS activation of the delay time averaged with only correct-WM tasks, correlated with WM performance in six channels (P < 0.05).

Subjects with better WM performance have higher levels of oxyhemoglobin activation compared with control trials in the WM delay time, and our results suggest that NIRS will be useful for measuring the WM performance.