To determine the prevalence and clinical correlations of an anomaly consisting of electroencephalographic (EEG) waves within the alpha frequency band during non-rapid eye movement (NREM) sleep in patients with fibromyalgia, and to evaluate the alpha NREM sleep anomaly as a predictor of response to amitriptyline. Twenty-two patients with fibromyalgia were studied in a 2-month, double-blind, crossover trial of amitriptyline (25 mg/day) versus placebo. Nocturnal EEGs were conducted on 2 consecutive nights at baseline and at the end of each 2-month treatment period. Six patients (27%) had a clinical response to amitriptyline, while none responded to placebo (P = 0.02). Treatment with amitriptyline or placebo did not result in any changes in the alpha ratings during NREM sleep. Only 8 patients (36%) exhibited the alpha NREM sleep anomaly at baseline. Those patients reported more sleep difficulty, but otherwise were clinically indistinguishable from those without this EEG sleep anomaly. Lower baseline alpha NREM sleep ratings were seen in responders to amitriptyline than in nonresponders, but these differences did not reach statistical significance. The alpha NREM sleep anomaly is present in only a small proportion of patients with fibromyalgia. It does not correlate with disease severity nor is it affected by treatment with amitriptyline. A larger sample size will be needed to adequately assess the value of this sleep anomaly in predicting the response to amitriptyline.