To examine basal ganglia dysfunction and atrophy in patients with mild to moderate Huntington's disease, with correlation of imaging measures with clinical and neuropsychological measures. Survey study in patients with Huntington's disease and matched controls, with imaging measures being evaluated by investigators unaware of the diagnosis. Baltimore Huntington's Disease Project, The Johns Hopkins Hospital, Baltimore, Md. Subjects included 10 patients with mild to moderate Huntington's disease and nine healthy age-matched control subjects. Imaging measures included single photon emission computed tomographic regional cerebral blood flow in caudate, putamen, and thalamus, and magnetic resonance imaging measures of caudate and putamen volumes and bicaudate ratios. Patients underwent neurologic and mental status examinations and neuropsychological tests. The measure with the greatest difference between patients and control subjects was mean putamen volume, reduced 54.3% in patients, with no overlap between groups (P<.001). Of the cerebral blood flow measures, caudate showed the greatest difference (21.5% decrease; P<.001). Quantitative neurologic indexes of disease severity correlated with both putamen measures (P<.03), while Mini-Mental State Examination scores correlated with caudate volume (P<.02). Bicaudate ratio correlated with both clinical measures and was the best index of neurologic deterioration (r=.95; P<.001), while global atrophy (measured by cerebrospinal fluid percentage) was the best correlate of several neuropsychological tests, such as the Trail Making Test (r=93; P<.001). Volumetric measurement of putamen best discriminated patients with Huntington's disease from healthy subjects. Measures of caudate atrophy or single photon emission computed tomographic measures performed less well. Neurologic decline correlated best with subcortical atrophy measured by the bicaudate ratio, but neuropsychological performance best corresponded to cerebrospinal fluid percentage, a measure of global atrophy.