You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster A4, Wednesday, November 8, 10:30 – 11:45 am, Harborview and Loch Raven Ballrooms

Depression alters limbic-sensorimotor brain interactions during implicit emotional speech production

Kevin Sitek1,2, Gregory Ciccarelli1,3, Mathias Goncalves1, Thomas Quatieri1,3, Satrajit Ghosh1,2;1MIT, 2Harvard University, 3MIT Lincoln Laboratory

Speech communication requires the precise coordination of language, motor, and sensory processes in the brain. Because speech is largely used to communicate thoughts, feelings, and internal states, additional cognitive and affective brain networks are likely involved in producing meaningful speech. We investigated the interaction between affect and speech production in participants with major depressive disorder (MDD). MDD is associated with emotional processing difficulties. In addition, depressed episodes affect speech production in individuals with MDD. Due to altered affective processing and speech production in depression, we hypothesized that sensorimotor and limbic brain activation would differ between MDD and non-depressed participants when producing implicitly emotional speech. Twenty-three participants with MDD and twenty-two non-depressed controls participated in an MRI experiment at MIT which included T1-weighted anatomical scans, diffusion-weighted imaging, resting-state functional imaging, and task functional imaging. For task fMRI, five slices were acquired simultaneously (SMS-5), allowing for quick whole-brain coverage at 2 mm isotropic resolution. A rapid sparse acquisition scheme alternated each 1.1 s acquisition with a 2.9 s silent gap. During these runs, participants saw a short sentence on the screen, which they then produced out loud during the silent period between acquisitions. Sentences were either sad, happy, or neutral, as rated by an independent sample, and were ordered pseudorandomly. Participants were not instructed to produce the sentences in any particular way with any explicit emotional prosody. Participants’ speech was recorded with an MRI-compatible microphone. Functional imaging data were preprocessed with Nipype. Overall, implicitly emotional speech production activated the canonical speech network, as well as medial prefrontal limbic regions, more strongly than neutral speech. However, the speech network was hyperactivated in MDD participants. Within the MDD group, depression severity was correlated with right insular activity when producing implicitly sad (compared to neutral) sentences. In resting state functional connectivity analysis, this right insular region was less connected with the speech production network but more connected with the default mode network in depressed participants compared to controls. Using psychophysiological interactions (PPI), we investigated functional connectivity between brain regions while performing the speech production task. We found that the amygdala, a structure involved in affective processing, was more functionally connected with right hemisphere insular and auditory cortices as well as regions associated with the default mode network in depressed (vs. control) participants. Compared to neutral speech, sad sentences showed less functional connectivity between the amygdala and posterior cingulate, another default mode region. In this study, we found that speech and limbic brain networks are activated differently between participants with MDD and non-depressed controls. In particular, the right insula, which has been implicated in both emotional processing dysfunction in depression as well as speech processing, is hyperactivated in the most depressed participants during implicitly sad speech and is functionally more strongly connected with both speech-motor and default mode regions. The right insula may play a crucial role in speech production changes that occur in depression.

Topic Area: Control, Selection, and Executive Processes

Back to Poster Schedule