Abstract

Despite a long history, the investigation of clinical depression using task-related, cognitive event-related potentials (ERPs) and quantitative EEG measures at rest have produced two distinct lines of research with little, if any, overlap in theory or findings. Whereas both literatures are typically limited by uncertainties attributable to volume conduction, a surface Laplacian or CSD transformation provides reference-free ERP/EEG measures that are closer to the underlying neuronal generators. Using a CSD-PCA approach, we recently reported a sharply localized novelty vertex source (NVS) component that is unique to novel stimuli and substantially reduced in depressed patients (Tenke et al 2010). In another study, condition-dependent (eyes open vs. closed) posterior EEG alpha predicted antidepressant treatment response (Tenke et al 2011). Inasmuch as the novelty ERP response is itself associated with alpha- and theta-coupling to slower rhythms (delta), we explored whether CSD-based TF measures could be linked to the NVS. Using 67-channel data from 49 depressed patients and 49 healthy adults, individual EEG epochs underlying target, nontarget, and novel ERP waveforms were: 1) CSD-transformed; 2) processed at each electrode site as event-related spectral perturbations (ERSP; limited to <400 ms and 30 Hz); and 3) simplified and quantified by unrestricted PCA. The most prominent TF factor (T273-F04: 273 ms peak latency at 4 Hz; 19% total variance) had a loadings peak latency comparable to the NVS (265 ms), but its midline topography was further anterior, pronounced for targets, and not reduced in patients. A subsequent component (T378-F10; 15%) had a posterior topography identifiable as posterior alpha desynchronization for both novels and targets that was greatest over the left hemisphere for controls (particularly for targets), but markedly reduced in patients compared to controls. However, an earlier midline component (T167-F10; 4%) represented alpha synchronization that was localized to the vertex, selective for novels, and larger for patients. Although this TF component corresponded to the mismatch negativity (MMN) latency peak to novels (175 ms), the MMN did not differ between groups. Results suggest that the enhanced early vertex synchronization of alpha observed in depressed patients following salient, but task-irrelevant, distracters may result in both the observed NVS reduction and the subsequent reduction of posterior condition-dependent alpha desynchronization in their response to novel sounds. These findings thereby provide a common framework for interpreting tonic and event-related alpha abnormalities in major depression.

Supported by NIMH grants MH094356 and MH36295.