Disentangling Clinical Profiles of Apathy in Behavioral Variant Frontotemporal Dementia

Abstract

Background: Apathy is highly frequent in behavioral variant frontotemporal dementia (bvFTD). It is presumed to involve different pathophysiological mechanisms and neuroanatomical regions.

Objective: We explored the hypothesis that subgroups showing distinct profiles of apathy and distinct patterns of atrophy within frontal lobes could be disentangled in bvFTD.

Methods: Using data-driven clustering applied to 20 bvFTD patients, we isolated subgroups according to their profiles on the three subscales of the Dimensional Apathy Scale (DAS). We explored their apathy profiles and atrophy patterns. Apathy profiles were characterized through both subjective measures of apathy by questionnaires and measures including objective behavioral metrics. Atrophy patterns were obtained by voxel-based morphometry, contrasting each bvFTD subgroup with healthy controls (N = 16).

Results: By clustering based on DAS dimensions, we disentangled three subgroups of bvFTD patients, with distinct apathy profiles and atrophy patterns. One subgroup, which presented the smallest pattern of atrophy (including orbitofrontal cortex) with a right asymmetry, was characterized by high self-reported emotional and initiation apathy and by a self-initiation deficit reversible by external guidance. In other subgroups showing more diffuse bilateral atrophies extending to lateral prefrontal cortex, apathy was not reversible by external guidance and more difficulty to focus on goal-management was observed, especially in the subgroup with the largest atrophy and highest levels of executive apathy.

Conclusion: Distinct clinical profiles of apathy, corresponding to distinct anatomical subtypes of bvFTD, were identified. These findings have implications for clinicians in a perspective of precision medicine as they could contribute to personalize treatments of apathy.

Keywords: Apathy; apathy subtypes; exploratory clustering; frontotemporal dementia; grey matter atrophy; voxel-based morphometry.

Fig. 1

Summary of data flow to manage measures of apathy profile and hypotheses on their conceptual validity. SAS, Starkstein Apathy Scale; DAS, Dimensional Apathy Scale; F1 and F2 are the two factors extracted by factor analysis; Questio_time_ratio: ratio of time spent exclusively for goal-directed actions related to the completing of a questionnaire in the guided phase.

Fig. 2

Three bvFTD subgroups and their profiles of apathy on DAS subscales. BvFTD-G1 in red (N = 8), bvFTD-G2 in green (N = 5), bvFTD-G3 in blue (N = 7). A) Results of the k-means clustering analysis (based on DAS subscales) defining three subgroups of bvFTD patients (each point with a number represents a bvFTD patient); B) Diminished integration, processing and expression of emotions (assessed by DAS-Emotional) in the three bvFTD subgroups; C) Lessened initiation of thoughts and actions (assessed by DAS-Initiation) in the three bvFTD subgroups; D) Inability to manage goals and cognitively strategize to execute a plan of actions (assessed by DAS-Executive) in the three bvFTD subgroups. Levels of significance (adjusted p-values): ns, non-significant; ^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

Fig. 3

Results supporting the validity of Questio_time_ratio as an indicator of ability to maintain focus on goal management in bvFTD patients. On the left: R indicates the Spearman’s rank correlation coefficient between Hayling error score and Questio_time_ratio. On the right: ß values are the standardized coefficients of the multiple linear regression analysis (with their associated p-values).

Fig. 4

Three bvFTD subgroups and their profiles of apathy on measures including behavioral metrics. bvFTD-G1 in red (N = 8), bvFTD-G2 in green (N = 5), bvFTD-G3 in blue (N = 7), and HC in yellow (N = 16). A) Results of the k-means clustering analysis (based on DAS subscales) defining three subgroups of bvFTD patients; B) Apathy defined as a global reduction of goal-directed behaviors (assessed by F1) in the three bvFTD subgroups compared to HC: no difference was observed between subgroups; C) Specific self-initiation deficit potentially reversible by hetero-guidance (assessed by F2) in the three bvFTD subgroups; D) Ability to focus on goal management (assessed by Questio-time-ratio) in the three bvFTD subgroups. Levels of significance (adjusted p-values): ns: non-significant; ^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

Fig. 5

Three bvFTD subgroups and their VBM– derived GM atrophy maps. bvFTD-G1 (N = 8), bvFTD-G2 (N = 5), and bvFTD-G3 (N = 7). The (1-p) value maps show the atrophy patterns compared with HC (N = 16) and are superimposed onto a whole-brain MNI template. Effects were corrected for age and sex, and statistical significance was set at p < 0.05 FWE-corrected for multiple comparisons.

Fig. 6

Summarized characteristics of the three bvFTD subgroups identified by data-driven clustering. DAS, Dimensional Apathy Scale; FAB, Frontal Assessment Battery; GDB, goal-directed behaviors; OFC, orbitofrontal cortex; ACC, anterior cingulate cortex; PFC, prefrontal cortex.