Clinical trials of cholinesterase inhibitor (ChEI) drugs, although generally reporting only minimal improvements in patients with Alzheimers disease (AD), indicate that a subgroup of patients may respond substantially to treatment. This study aimed to assess the clinically variable ChEI treatment effects in a group of patients with mild AD using a semantic association and an N-back light working memory activation paradigm. Twenty-six patients with probable mild AD treated with a ChEI for 20 weeks were retrospectively divided into responders and non-responders. Patients were classified as responders if their Clinicians Interview Based Impression of Change (CIBIC - Plus) score was four or less and if they had an increase of at least two points on the MMSE. These criteria resulted in two subgroups comprising nine responders and seventeen non responders. Nine healthy elderly age-matched controls were also recruited as a comparison group. ChEI treatment was accompanied by significant modulation of task induced activation increases in both fMRI tasks in AD responders. The effect of ChEI response was in effect a restoration of regional brain function in the same areas used by elderly controls when performing these tasks. In non-responders decrements in task related activation were observed and over time activation patterns appeared less like the elderly controls. Screening semantic fluency scores correlated negatively with activation increases at retest. In the paper, a tentative explanation is offered of why subgroups of patients with a similar clinical diagnosis and level of clinical severity show a different physiological response to ChE inhibition.