Abstract

Cognitive behavioral therapy (CBT-I) is a common and effective method for treating chronic insomnia, although patient responses to it are not uniform and research on predicting treatment response has mostly focused on psychological factors. Here, it is investigated whether brain oscillations during sleep at baseline, particularly sleep spindles, are predictive of treatment response. Twenty-four participants with chronic primary insomnia took part in a 6-week CBT-I performed in groups of 4 to 6 participants. Treatment response to CBT-I was assessed using the Pittsburgh Sleep Quality Index (PSQI) and the Insomnia Severity Index (ISI) measured at pre- and post-treatment. Secondary outcome measures included sleep diary (over seven days) and polysomnography (PSG) sleep efficiency (%) measured at pre- and post-treatment. Spindle density (as well as secondary measures of duration, amplitude, power, frequency, and spectral power in the sigma band) during stages N2-N3 sleep were extracted from the PSG recording at pre-treatment. Multiple regression assessed whether sleep spindle activity predicted treatment response to CBT-I. After controlling for baseline measures, age, sex, education level, treatment compliance, time in N2, and the location of the sleep recording, lower spindle density and sigma power at pre-treatment predicted poorer CBT-I response at post-treatment, as reflected by lower PSQI scores.