Blame-rebalance fMRI neurofeedback in major depressive disorder: A randomised proof-of-concept trial

Abstract

Previously, using fMRI, we demonstrated lower connectivity between right anterior superior temporal (ATL) and anterior subgenual cingulate (SCC) regions while patients with major depressive disorder (MDD) experience guilt. This neural signature was detected despite symptomatic remission which suggested a putative role in vulnerability. This randomised controlled double-blind parallel group clinical trial investigated whether patients with MDD are able to voluntarily modulate this neural signature. To this end, we developed a fMRI neurofeedback software (FRIEND), which measures ATL-SCC coupling and displays its levels in real time. Twenty-eight patients with remitted MDD were randomised to two groups, each receiving one session of fMRI neurofeedback whilst retrieving guilt and indignation/anger-related autobiographical memories. They were instructed to feel the emotion whilst trying to increase the level of a thermometer-like display on a screen. Active intervention group: The thermometer levels increased with increasing levels of ATL-SCC correlations in the guilt condition. Control intervention group: The thermometer levels decreased when correlation levels deviated from the previous baseline level in the guilt condition, thus reinforcing stable correlations. Both groups also received feedback during the indignation condition reinforcing stable correlations. We confirmed our predictions that patients in the active intervention group were indeed able to increase levels of ATL-SCC correlations for guilt vs. indignation and their self-esteem after training compared to before training and that this differed significantly from the control intervention group. These data provide proof-of-concept for a novel treatment target for MDD patients and are in keeping with the hypothesis that ATL-SCC connectivity plays a key role in self-worth.https://clinicaltrials.gov/ct2/show/results/NCT01920490

Reinforcement and Punishment Shape the Learning Dynamics in fMRI Neurofeedback

“…The decrease in PANAS-SF score, which was intended to capture effects of the NF training on emotions, can be explained by the demanding task but also subjectively relaxing scanner environment (as reported during the debriefing) in conjunction with the focus on activity of the positively valenced items. Similar NF-related decreases on the positive affect scale of the PANAS have been previously reported ( Lorenzetti et al, 2018 ; Zahn et al, 2019 ). It has further been shown that NF has a significant influence on feelings of fatigue ( Rana et al, 2016 ) which, in turn, might affect the regulation performance ( Moll et al, 2014 ).…”

A systematic review of fMRI neurofeedback reporting and effects in clinical populations

“… – n.a. – – – – Mehler et al, 2018 @ Depression 16 Mainly anterior brain areas such as insula and striatum Sug Yes ✗ * Clin Diff str ✓ ↑ Yes No Sess Clin ↑ – 18w (clin ↑) NCT01544205 Jaeckle et al, 2019 ^@$ Depression 19 rSATL and pSCC No No * Clin Diff str ✓ ↓ Yes – Yes Clin ↑ No – ISRCTN10526888 Zahn et al, 2019 $ Depression 2 14 aSCC and aSTC No No ** Clin Diff con ✗ ↑ Yes Yes Pre-post Clin – – – NCT01920490 Rance et al, 2018 Misc 10 (A), 20 (TS) 5 OFC (A), SMA (TS) Yes – ** Clin Yoked – ↓↑ same – – – Clin ↑ – 2, 4, 6, 8w (clin ↑) NCT02206945 NCT01702077 McDonald et al, 2017 Misc 76 6 DMN Yes No – ↓↑ same Yes – – (n.a.) – – – – Skouras and Scharnowski, 2019 @ Misc 74 DMN Yes ...…”

Self-blaming emotions in major depression: a randomised pilot trial comparing fMRI neurofeedback training with self-guided psychological strategies (NeuroMooD)

“…Group level analyses of primary and secondary outcomes, comparing pre-and post-treatment effects (visit 1 vs visit 5), were obtained using the constrained longitudinal analysis model (cLDA; Coffman, Edelman & Woolson, 2016) after seeking statistical advice (K.G., E.C. ), the alpha-level was set to p=.05, two-tailed.As in our previous paper(Zahn et al, 2018) at the individual subject level, linear regression coefficients for the slope of z-transformed ATL signal timecourse as the predictor of z-transformed SC signal time-course in each condition (self-blame, indignation/anger) in the pre-and post-training acquisition as the outcome variables were derived from a general linear model for each subject by modelling the interaction of z-transformed ATL signal time-course with two factors: condition (guilt, indignation) and time (pre-, post-training). The ztransformation was undertaken to obtain standardised regression coefficients.Cohen's d effect sizes were computed for each regression coefficient using the formula: 2 x t-value/square root of degrees of freedom (df;Rosenthal & Rosnow, 1991).Where cLDA was inapplicable, intervention group comparisons were performed using non-parametric tests.…”

“…FRIEND has previously been validated for correlation feedback in patients with MDD (Zahn et al, 2018).…”

“…The first five volumes of each emotional block were discarded due to high correlations guided by a decrease in time series after subtraction conditions (Zahn et al, 2018).…”