Quantification of occipital EEG changes in Alzheimer's disease utilizing a new metric: the fractal dimension.

In a retrospective preliminary investigation, the occipital electroencephalogram (EEG) changes associated with Alzheimer's disease (AD) were examined using a novel quantification metric: the fractal dimension. The mean occipital EEG fractal dimension was determined for each of three patient groups representing a spectrum of clinical and EEG pathology: controls, probable AD, and autopsy-confirmed AD. The fractal dimension was significantly reduced in each of the AD groups with respect to controls (p < 0.001); and within the AD groups, it was significantly reduced in autopsy-confirmed AD relative to probable AD (p < 0.01). The fractal dimension findings paralleled the manifest EEG abnormalities in a way that suggests it has potential clinical utility in metric studies on the EEG, especially when applied to the dementias. Additionally, the EEG pathology studied here was particularly well-described by the fractal dimension, providing further support for a nonlinear approach to the background activity of the EEG.