Effects of different language and tDCS interventions in PPA and their neural correlates

Language intervention has been shown to critically affect the course of post-stroke language rehabilitation. Interventions at early or chronic stages (either by reperfusion, speech-language therapy or tDCS) have dramatically improved outcomes in language post-stroke rehabilitation. Little is known, however, about interventions in neurodegerative diseases and the changes in neural substrates they may induce. It has been shown that language therapy is indeed beneficial and improves language outcomes, both with behavioral interventions and more recently with tDCS for longer lasting and more generalizable effects even in PPA(Cotelli et al., 2014; Tsapkini, Frangakis, Gomez, Davis, & Hillis, 2014). However, it has not been addressed how this is instantiated in the brain. In this paper we evaluate the effects of tDCS in 18 participants with PPA who received both sham and tDCS coupled with language therapy. Our primary aim was to evaluate the effects of intervention and determine the neural correlates of tDCS vs. sham interventions that led to improved language outcomes using resting-state fMRI (rs-fMRI). Method: Eighteen patients diagnosed with PPA underwent written or oral language production intervention with and without tDCS (sham) in a within-subjects cross-over design. Participants received treatment for 2 weeks, 10 sessions for each condition. Each condition was separated by 3 months. Resting-state fMRI data were obtained on the first treatment condition from 13 participants who underwent written language intervention at 3 time-points: before, after and 2-months post-intervention. Participants were pseudo-randomly assigned in either tDCS or sham condition first in a between-subjects design. Results: First, we replicated our previous results obtained with fewer participants: all improved in both tDCS and sham conditions on trained items. Generalization of treatment on untrained items was significant only in tDCS condition. Therapy gains lasted longer in tDCS condition as well. Second, preliminary analyses of rs-fMRI show changes of functional connectivity between written language areas in the tDCS and sham conditions. Conclusions: tDCS represents an increasingly valuable treatment option in language rehabilitation even in neurodegeneration. Late intervention is as beneficial as early intervention but improvement seems more dramatic in early cases. Different possibilities are discussed: tDCS may indeed change the course of the disease, i.e., it may slow down the rate of decline or, language improvement due to tDCS (or delay in language deterioration due to the course of the disease) may hold the spread of decline in other cognitive functions, thus, early interventions appear more beneficial. The correlation between functional connectivity and language production outcomes is expected to shed light on how tDCS works in the brains of people with a neurodegenerative disease. Implications of functional connectivity changes between language areas involved in the targeted language function will inform further interventions.