Previous research using a conditioned-attention paradigm demonstrated that 4-month-old infants of depressed mothers (a) failed to acquire associations when a segment of their mothers' infant-directed (ID) speech signaled the presentation of a smiling face but (b) did acquire associations when a segment of an unfamiliar nondepressed mother's ID speech signaled the face (P. S. Kaplan, J. -A. Bachorowski, M. J. Smoski, & W. J. Hudenko, 2002). In the present study, 5- to 13-month-old infants of depressed mothers failed to acquire associations when either their own mothers' (Experiment 1) or an unfamiliar nondepressed mother's (Experiments 1 and 2) ID speech signaled a face. However, these infants acquired associations when a segment of an unfamiliar nondepressed father's ID speech served as the signal (Experiment 2). One possible explanation of these results is that infants of depressed mothers selectively "tune out" ID speech from their mothers and from other, nondepressed, women.