Abstract

Auditory EP loudness dependency (LDAEP) shows promise as a predictor of clinical response in depressed patients treated with serotonin agonists. Quantification of LDAEP used N1/P2 peak-to-peak differences and inverse models of N1 corresponding to primary auditory cortex generators (e.g., BESA, LORETA). Reference-free unrestricted Varimax CSD-PCA (covariance matrix) offers a conservative, model-independent alternative. Auditory 72-channel ERPs (BioSemi) were recorded from 23 healthy adults and 15 depressed patients listening to 40-ms, 1000-Hz pure binaural tones (1600-2100 ms ISI) at five, equiprobable intensities (60-100 dB SPL). Subsequently, patients began treatment with an SSRI, NDRI, or both antidepressants, and were assessed for treatment response after 4-12 weeks (7 remitters; 8 nonremitters). An N1 factor (116 ms peak latency) with a tangentially-oriented sink/source topography consistent with activation of auditory cortex showed a robust, monotonic association with intensity. Whereas the loudness dependency of this sink was greater for remitters than for nonremitters, who did not differ from controls, only overall N1 sink was significantly greater for remitters. The P2 factor (226 ms; midline source topography) showed intensity effects that were less robust, while a radial temporal lobe N1 sink (167 ms) and a temporal lobe P3 source (351 ms) failed to show consistent effects. Thus, CSD-PCA offers a concise, but conservative, characterization of LDAEP generators, and may be of clinical value for predicting response to antidepressants. [Supported by NIMH grants MH036295 and MH066597.]