The aim of this study was to investigate brain tissue volumes, grey matter (GM) distribution, and cognitive performance for cognitively impaired subjects using cerebrospinal fluid (CSF) biomarker cut-offs as grouping criteria. 41 subjects attending the Memory Clinic, Karolinska University Hospital, Huddinge, Sweden, were divided into groups based on normal or abnormal CSF levels of Aβ1-42, t-tau, and p-tau181. SIENAX algorithms were employed for brain tissue volumes estimation and voxel-based morphometry (VBM) for mapping the differences in GM patterns. VBM revealed significant lower GM volumes in temporo-parietal, occipital, and prefrontal cortices for those subjects belonging to abnormal CSF t-tau and p-tau181 groups. No differences were found between groups according to CSF Aβ1-42 cut-offs. Patients with abnormal CSF p-tau181 showed lower cognitive performance compared to those with normal levels. Patients with abnormal levels of CSF tau (but not Aβ1-42) showed an Alzheimer's disease-like pattern for both GM distribution and cognitive profile, compared to those with normal levels. These results support the hypothesis that CSF t-tau or p-tau181 levels may be of direct value for the evaluation of disease severity.