Aberrant functional connectivity in depression as an index of state and trait rumination

Depression has been shown to be related to a variety of aberrant brain functions and structures. Particularly the investigation of alterations in functional connectivity (FC) in major depressive disorder (MDD) has been a promising endeavor, since a better understanding of pathological brain networks may foster our understanding of the disease. However, the underling mechanisms of aberrant FC in MDD are largely unclear. Using functional near-infrared spectroscopy (fNIRS) we investigated FC in the cortical parts of the default mode network (DMN) during resting-state in patients with current MDD. Additionally, we used qualitative and quantitative measures of psychological processes (e.g., state/trait rumination, mind-wandering) to investigate their contribution to differences in FC between depressed and non-depressed subjects. Our results indicate that 40% of the patients report spontaneous rumination during resting-state. Depressed subjects showed reduced FC in parts of the DMN compared to healthy controls. This finding was linked to the process of state/trait rumination. While rumination was negatively correlated with FC in the cortical parts of the DMN, mind-wandering showed positive associations.

Mind-wandering: 1) I felt relaxed. 2) I let my mind flow. Rumination: 3) I ruminated (in the sense of revolving thoughts). 4) I thought about things I have to do/ tried to make plans. 5) I tried to fight certain experiences. 6) I felt stressed. Focus on sensations: 7) I felt body sensations.

8) I concentrated on things I hear. Fight against fatigue:
9) I thought about the duration of the measurement. 10) I needed to fight falling asleep.

Additional information on the rating of the self-report form
To validate the used VAS scales and for reasons of additional information on restingstate processes, we also used a qualitative self-report form. On a blank page, subjects were asked to note the experiences they had during the resting-state measurement.
The instruction was as follows: "Please describe in the following what you did during the resting state measurement and how you felt. You may answer the following questions: What did you feel and think during the measurement? How did you react to your thoughts and feelings? What consequences followed your reactions?" The texts were screened and categorized by two independent raters to assess qualitative measures of processes during resting-state according to qualitative methods: First, self-report forms were analyzed and categories were built and defined until saturation was reached. The following categories were defined: • Mind-wandering: The subject expressed to be in a relaxed mood and let his mind flow in an unconstrained way without any focus on a particular subject.
Example: "I relaxed and let my mind flow." Example: "I thought about things that matter to me, but I was not stuck in my thoughts. I liked to let my mind flow." • Rumination: The subject expressed a repetitive stressful style of thinking about an unfinished concern that leads to the urge of suppressing the inner experience.
Examples: "I thought about a stressful meeting I had at work, which made me nervous, so I tried to distract myself from that memory." "I thought about an argument with my boyfriend and asked myself what I am doing wrong." • Focus on body sensations: The subject expressed an attentional focus on their body.
Examples: "I focused on my breathing." "I felt my body and my heartbeat." • Mindfulness/Relaxation training: The subject expressed to be in a mindful state (detachment from cognition, concentration on breathing with detached mind) or to perform some kind of relaxation technique (e.g. progressive muscle relaxation).
Example: "I focused on my breathing and watched my mind in a detached way." • Suppression: The subject expressed withdrawal from or suppression of unpleasant inner experiences.
• Boredom: The subject expressed that the resting-state was boring.
• Unfinished business: The subject expressed thoughts about things they will do. Examples: "I thought about what I would eat for dinner and decided to eat pizza." "I thought about the homework I have to do." • Thinking about the measurement: The subject expressed thoughts about the given instructions or how their data might look like.
• Fight against fatigue: The subject expressed feeling sleepy or trying not to fall asleep.
• Thoughts about the duration of the measurement: The subject expressed thoughts about the duration of the measurement or counted the time.
Afterwards, the most common categories were used to categorize self-report forms by two independent psychologists.

Influence of cofounders
Regarding effects of other resting-state process variables, there was no effect for the factors "focus on sensations" and "fighting against fatigue". One reason for this finding may be that the variance for these scales was smaller, since many participants focused on body sensations and felt sleepy at some point of the resting-state measurement.
In contrast to that, the scale for measuring mind-wandering was positively associated with FC in the DMN, as expected (see supplemental material Figure S2). NBS analysis of the factor revealed a significant (p=.026±0.0045) network with 28 nodes and 39 edges, reflecting higher FC in participants reporting high mind-wandering (see supplemental material Figure S3 and Table S1).
Medication status had no effect on FC-differences between depressed medicated and depressed non-medicated subjects (p>.1).

FC properties in the probeset
For the whole sample, FC coefficients in the used probeset showed an expected distribution with high connectivity within DMN regions of the middle parietal cortex and the supramarginal gyrus (SupG) and angular gyrus (AngG). These regions showedas assumed -low to negative FC with the temporal cortex consisting of the superior temporal gyrus, fusiform gyrus and subcentral area (see figure S1). In addition, the self-reported amount of mind-wandering correlated positively with FC measures (see supplemental material Figure S2) and showed significant network differences between subjects reporting high vs. low mind-wandering within the DMN with hub nodes in the middle somatosensory cortex (SAC) and the SupG (see supplemental material: table S1 and Figure S3).  Figure S3. NBS analysis of the main effect for mind-wandering. Left: Significantly hyperconnected network for "high mind-wanderers". Right: FC maps for the contrast "high vs. low mind-wandering" in the seed region of the left supramarginal cortex. Results of the NBS analysis can be seen in table s1.    Table S4. Item characteristics of the resting-state VAS scales.

Supplementary Analysis:
As supplementary exploratory analysis, we performed two different analyses that were not directly related to the research question, but gave additional information for the interpretation of the results: 1) We performed a rumination subgroup analysis as defined by the qualitative rumination rating (based on the self-report form) in the MDD group only. To this end, the 40% of the MDD subjects reporting rumination in the self-report form during resting-state were compared with the 60% which did not report rumination in the self-report form.
2) In the main analysis, the RRS total score was used. In a third analysis, we also correlated sub-scores of the questionnaire (brooding and reflection) with the FC scores in the whole sample.
Supplementary Analysis of the qualitative rumination rating in the MDD group only.
The analysis of the 40% of the depressed subjects that reported rumination in the selfreport form as compared to the 60% of the depressed subjects that did not report rumination revealed a significant disconnected network (with 36 nodes and 67 edges, t(58)=2.7, p=.003). The network -which showed lower FC in high ruminating subjects -was bilaterally organized and had hubs in the bilateral fusiform gyri and somatosensory association cortex. However, inter-hemispheric disconnections were rare and mediated over central hubs. Effect sizes ranged between d=-.44 to d=-96 within FC to the seed channel in the somatosensory association cortex and between d=-.56 to d=-.94 in the right fusiform gyrus. Figure S6. Differences between the subgroup depressed high ruminators and depressed low ruminators according to the qualitative self-report forms. Blue colors indicate reduced FC in high ruminators.