the objective of the present study was to investigate within a prospective design the seasonality and depression predictors of light therapy response after controlling for treatment expectancy and pre-therapy functioning. seasonality was indicated by the global seasonality score (oss) on the seasonal pattern assessment questionnaire (spaq). the depression predictors were measured with the structured interview guide for the hamilton depression rating scale - seasonal affective disorder version (sigh-sad) that yielded severity levels for the typical and atypical depressive symptom. seventeen participants received 30 minutes of bright light therapy at 10,000 lux daily for 14 days. all participants had a screening sigh-sad score of at least 22, were not receiving antidepressant treatment at least 4 weeks (8 weeks for fluoxetine) before the study, had no previous experience with light therapy, were free of pre-existing retinal or eye diseases, and were free of use photosensitizing medications or supplements. their treatment expectancy, pre-therapy and post-therapy functioning on several variables were measured. significant increase was found in sleep quality and significant decrease was found in carbohydrate craving, carbohydrate intake, atypical symptom score, typical symptom score, and total depression score. higher expectations of treatment predicted lower carbohydrate craving following light therapy. higher typical symptom score before light therapy predicted higher atypical symptom score and higher total symptom score following light therapy. the gss bore no relationship to treatment outcome. the present study provides support for previous research findings that individuals with more severe typical depressive symptoms have a poorer response to light therapy. findings also underscore the need for placebo control in light therapy studies.