Subjective time perception in dementia: a behavioural and neuroanatomical analysis

Subjective time perception-the modulation of elapsed clock time by sensory, homeostatic and psychological factors-is fundamental to how we experience the world. People with Alzheimer's disease and frontotemporal dementia often exhibit clinically relevant symptoms of altered temporal awareness but these are poorly understood. Here we addressed this issue in a cross-sectional, case-control study of 60 patients representing all major Alzheimer (n = 24) and frontotemporal (n = 36) syndromes [mean age 68.8 (range 48-77 years); 28% female] and 24 cognitively well age-matched controls [age 69.4 (6.5) years; 50% female]. Subjective duration perception was assessed using an auditory temporal bisection paradigm, in which the task was to compare sound stimulus durations with learned (2 and 5 s) reference intervals. We varied sound emotional valence and semantic identity (behavioural salience) to create four stimulus conditions: pleasant environmental (running water), unpleasant environmental (machine noise), pleasant human (laughter) and unpleasant human (crying) sounds. Psychometric functions were constructed to assess sound duration estimation (bisection point) and sensitivity (Weber's ratio), and participant groups were compared using linear mixed regression models. Neuroanatomical associations of altered subjective time perception (sound duration estimation) were assessed using voxel-based morphometry of patients' brain MRI images. All participants perceived environmental sounds as lasting longer than human sounds, unpleasant environmental sounds as longer than pleasant environmental sounds and pleasant human sounds as longer than unpleasant human sounds (all P < 0.05). In dementia syndromes, the effect of sound semantic category was accentuated: patients with nonfluent variant primary progressive aphasia overestimated environmental sound duration, while patients with logopenic aphasia underestimated the duration of human sounds, relative to controls (P < 0.05). In addition, patients with typical Alzheimer's disease and behavioural variant frontotemporal dementia discriminated sound duration changes less sensitively than controls, while patients with semantic variant primary progressive aphasia discriminated sound duration more sensitively than other syndromic groups (P < 0.05). Neuroanatomical correlates of auditory duration perception were identified for different sound categories, in distributed cortical areas previously implicated in the pathogenesis of these diseases (all significant at P < 0.05, after correction for multiple voxel-wise comparisons in pre-specified regions of interest): precuneus (environmental sounds), supramarginal gyrus (pleasant human sounds) and insula (unpleasant human sounds). Our findings show that canonical dementia syndromes have clinical and neuroanatomical signatures of altered subjective time perception, linked to clinically relevant properties of sensory stimuli and the core pathophysiology of frontotemporal dementia and Alzheimer's disease. The findings suggest a novel paradigm for characterizing these diseases, with diagnostic and management implications.