Dynamic network characteristics of adolescents with major depressive disorder: Attention network mediates the association between anhedonia and attentional deficit

FIGURE 1. Schematic diagram of experimental process.

FIGURE 2. Intrinsic networks (ICNs) are derived from group‐independent component analysis. Maps were thresholded at p uncorrected < .001. The frontoparietal network (FPN) includes the prefrontal cortex and posterior cingulate cortex (PCC); the default mode network (DMN) includes medial prefrontal cortex (mPFC), lateral parietal (LP), and PCC; the dorsal attention network (DAN) includes the frontal eye field (FEF) and intraparietal sulcus (IPS).

FIGURE 3. (a) Electroencephalography microstate topology. (b) Comparison of EEG microstate indicators between groups in the baseline. Light color: MDD group (n = 188); dark color: HC group (n = 94). (A) Mean duration (ms) of each EEG microstate in the two groups. (B) Frequency of each EEG microstate in the two groups (%). (C) Time ratio of each EEG microstate in the two groups (%). (D) State transition probability between pairwise EEG microstates in the two groups (%). MDD, major depression disorder; HC, healthy controls; *p < .05; **p < .01. (c) Correlation analysis of microstate with clinical symptoms and cognitive function and mediating effect analysis. (A) Coverage of Microstate C was negatively correlated with the HAMD score of the adolescent MDD group. (B) Coverage of Microstate D was significantly positively correlated with attention function. (C) The frequency of Microstate D was significantly positively correlated with immediate memory. (D) Duration of Microstate A was significantly positively correlated with immediate memory. (E–G) Correlation analysis showed that the transition probability from Microstate C to Microstate D was significantly positively correlated with anhedonia…

FIGURE 4. Comparison of EEG microstate indicators between groups in 8 weeks. White: MDD group (n = 188); black: HC group (n = 94). (a) Mean duration (ms) of each EEG microstate in the two groups. (b) Frequency of each EEG microstate in the two groups (%). (c) Time ratio of each EEG microstate in the two groups (%). (d) State transition probability between pairwise EEG microstates in the two groups (%). RP, remission prediction; NRP, non‐remission prediction group.

FIGURE 5. Temporal properties of FPN. Scatter plot of temporal frequency for each participant (a(A)). There was no significant difference between the two groups in temporal frequency of FPN (a(B)). Scatter plot of temporal variability for each participant (b(A)). There was no significant difference between the two groups in temporal variability of FPN (b(B)). Correlation analyses indicated that temporal variability of FPN was not related to either anhedonia or attention in MDD patients (c). HC, healthy controls; MDD, major depression disorder; ns, no significance.

FIGURE 6. Temporal properties of DMN. Scatter plot of temporal frequency for each participant (a(A)). There was no significant difference between the two groups in temporal frequency of DMN (a(B)). Scatter plot of temporal variability for each participant (b(A)). There was no significant difference between the two groups in temporal variability of DMN (b(B)). Correlation analyses indicated that temporal variability of DMN was positively correlated with attention scores (c(B)), while not related with anhedonia scores in MDD patients (c(A)). HC, healthy controls; MDD, major depression disorder; ns, no significance.

FIGURE 7. Temporal properties of DAN. Scatter plot of temporal frequency for each participant (a(A)). There was no significant difference between the two groups in the temporal frequency of DAN (a(B)). Scatter plot of temporal variability for each participant (b(A)). MDD patients showed significantly weaker temporal variability of DAN (b(B)) than healthy controls. Correlation analyses indicated that the temporal variability was significantly positively correlated with anhedonia and negatively related with attention scores in MDD patients (c(A,B)). Mediation analysis revealed that the temporal variability of DAN mediated the relationship from anhedonia to attention (c(C)). HC, healthy controls; MDD, major depression disorder; *p < .05; ns, no significance.

FIGURE 8. Temporal properties of DAN after 8 weeks' treatment. Scatter plot of temporal frequency for each participant (a(A)). There was no significant difference between the remitters and treatment failures in the temporal frequency of DAN (a(B)). Scatter plot of temporal variability for each participant (b(A)). Remitters showed significantly higher temporal variability of DAN (b(B)) than treatment failures. MDDR, remitters; MDDF, treatment failures; *p < .05; ns, no significance.