Non-invasive measures of brain gamma-aminobutyric acid (GABA) concentrations may be especially useful in the identification of cocaine-related changes in brain chemistry that can be used to guide the development of future treatments for cocaine-dependent persons. This study assessed whether brain GABA levels in cocaine-dependent subjects with and without an alcohol disorder differ from GABA levels in healthy comparison subjects. Two-dimensional, proton magnetic resonance spectroscopy was used to determine GABA levels in the left prefrontal lobe of cocaine-dependent subjects (N=35) recruited from a National Institute on Drug Abuse (NIDA)-sponsored treatment trial of cocaine dependence and a comparison group (N=20). At treatment baseline, mean GABA concentrations were 0.93+/-0.27 mM/kg in cocaine-dependent subjects and 1.32+/-0.44 mM/kg in the comparison sample (t [d.f.=53]=3.65, P<0.001). Cocaine-dependent subjects with a history of a co-morbid alcohol disorder (N=23) had significantly lower baseline GABA levels (0.87 mM/kg) (t [d.f.=41]=4.31, P<0.001) than the comparison group. However, cocaine-dependent subjects without an alcohol disorder (N=12) also had lower GABA levels (1.04 mM/kg) than the comparison subjects (t [d.f.=30]=2.09, P=0.045), suggesting that cocaine dependence alone can decrease GABA levels.