Evaluating state-based network dynamics in anhedonia

Abstract

Anhedonia is a transdiagnostic clinical syndrome associated with significant clinical impairment. In spite of this, a clear network-level characterization of anhedonia does not yet exist. The present study addressed this gap in the literature by taking a graph theoretical approach to characterizing state-based (i.e., reward anticipation, rest) network dynamics in a transdiagnostic sample of adults with clinically significant anhedonia (n = 77). Analyses focused on three canonical brain networks: the Salience Network (SN), the Default Mode Network (DMN) and the Central Executive Network (CEN), with hypotheses focusing on the role of saliency-mapping in anhedonia. Contrary to hypotheses, no significant relation was found between the SN and anhedonia symptom severity. Exploratory results revealed a significant association between anhedonia severity and DMN reorganization from rest to reward anticipation. Specifically, greater anhedonia severity was associated with less reward-related reorganization. This finding suggests that anhedonia severity may be associated with DMN hyposensitivity, such that individuals with more severe anhedonia may have a difficult time disengaging from their internal world in the context of potentially rewarding experiences. Although preliminary, this finding challenges the centrality of the SN in anhedonia severity and suggests the importance of the DMN. Clinical implications and future directions are explored.

Keywords: Anhedonia; Default mode network; Graph theory; Network dynamics; Reward; Salience network.

Fig. 1

Monetary incentive delay task. Note. ISI = Inter-Stimulus Interval; ITI = Inter-Trial Interval. Cue = fixed: 500 ms; ISI 1 = jittered: 2250–3750 ms; Target = fixed: 150 ms; ISI 2 = jittered: 2400–2750 ms; Feedback = fixed: 1250 ms; ITI = jittered to fit task length.

Fig. 2

Network visualization.

Fig. 3

Conceptual depiction of network reorganization.

Fig. 4

Reward anticipation > neutral anticipation. Note. Group-level activation of reward anticipation > neutral anticipation. Z-statistic image thresholded non-parametrically using clusters determined by Z > 2.3, cluster significant threshold of p = 0.05.

Fig. 5

Neutral vs. Reward Anticipation Activation Across Networks. Note. Task-based activation averaged across ROIs within each network. ∗ = p < 0.05.

Fig. 6

Between-network connectivity. Note. Correlation coefficients denoting between-network connectivity. Reward condition = bold, neutral condition = regular, rest = italicized text. No significant state-based differences. No significant reward-related differences (neutral vs. reward), but significant decrease in between-network connectivity from rest to task (both neutral and reward), p < 0.05.

Fig. 7

Within-network connectivity. Note. Correlation coefficients denoting within-network connectivity. Reward condition = bold, neutral condition = regular, rest = italicized text. No significant reward-related differences (neutral vs. reward), but significant decrease in within-network connectivity from rest to task (both neutral and reward), p < 0.05.

Fig. 8

Reorganization between rest and reward anticipation.

Fig. 9

DMN reorganization and anhedonia. Note. Significant correlation between anhedonia severity and reorganization within the DMN when transitioning between rest and reward anticipation and anhedonia severity.

Fig. 10

DMN within-network connectivity.