Background Functional cognitive disorders (FCD) are an important differential diagnosis of neurodegenerative disease. The utility of suggested diagnostic features has not been prospectively explored in 'real world' clinical populations. This study aimed to identify positive clinical markers of FCD. Methods Adults with cognitive complaints but not dementia were recruited from memory, neurology, and neuropsychiatry clinics. Participants underwent structured interview, MINI, MoCA, Luria 3-step, interlocking fingers, digit span and MSVT, PHQ-15, HADS, MMQ, and PSQI. Potential diagnostic variables were tested against expert consensus diagnosis using logistic regression. Results FCD were identified in 31/49 participants. Participants with FCD were younger, spoke for longer when prompted 'Tell me about the problems you've been having', and had more anxiety and depression symptoms and psychiatric diagnoses than those without FCD. There were no significant differences in sex, education, or cognitive scores. Younger age and longer spoken response predicted FCD diagnosis in a model which explained 74% of diagnostic variability and had an AUC of 94%. Conclusions A detailed description of cognitive failure is a sensitive and specific positive feature of FCD, demonstrating internal inconsistency between experienced and observed function. Cognitive and performance validity tests appear less helpful in FCD diagnosis. People with FCD are not 'worried well' but often perform poorly on tests, and have more anxiety, depression, and physical symptoms than people with other cognitive disorders. Identifying diagnostic profiles is an important step towards parity of esteem for FCDs, as differential diagnoses of neurodegenerative disease and an independent target for clinical trials.