EEG spectral changes in treatment-naive, actively drinking alcoholics

Abstract

Background: The current study examines the EEG spectra of actively drinking, treatment-naive alcoholics (TxNAs).

Methods: EEGs were gathered on 51 TxNAs and age- and sex-matched controls during closed-eyes conditions. Participants were excluded for lifetime diagnoses of psychiatric or substance abuse disorders. Power for the theta to high beta bands was examined across midline electrodes.

Results: The TxNA sample exhibited a nexus of disinhibited traits associated with the vulnerability to alcoholism and had developed alcohol dependence but no other diagnosable psychiatric or substance abuse disorders. The TxNAs evidenced higher power for all EEG bands compared with controls. The magnitude and anterior-posterior extent of the group differences varied across bands. Within TxNA, EEG power was negatively correlated with average and peak alcohol drinking duration dose.

Conclusions: Increased EEG power across the theta to high beta bands distinguishes TxNAs without comorbid diagnoses from controls. These effects varied across bands in magnitude and spatial extent, suggesting different effects for the different EEG spectral generators. The authors hypothesize that the increased power in these individuals is a trait difference associated with the inherited nexus of disinhibited traits and its manifestation in alcoholism. Based on the strong negative correlations with alcohol use variables, the authors speculate that decreases in EEG power are a morbid effect of long-term alcohol abuse. They acknowledge that this hypothesized effect of alcohol abuse on EEG power is opposite to the increased EEG power that they hypothesize is associated with alcoholism and its inherited nexus of disinhibited traits. An implication of this model is that with continuing alcohol abuse, the increased EEG power in TxNAs will eventually be overpowered by the effects of long-term severe alcohol abuse. This model predicts that in very long-term alcoholics, EEG power would be equal to or lower than that of age- and sex-comparable controls.

Fig. 1

Displays group differences in EEG power for each band at each midline electrode location. For presentation purposes the inverse of log (power) has been used to show the results as power on a natural log scale. *, **, ***: p < .05, p < .005, p < .0001, p values of trends are noted.