Slide Sessions

Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams

Slide Session C: Disorders

Saturday, October 8, 10:30 am - 12:00 pm EDT, Regency Ballroom

Chair: William Matchin, University of South Carolina

Effects of transcranial alternating current stimulation on language fluency in post-stroke aphasia: A proof-of-concept study

Lynsey Keator¹, Lisa Johnson, Julius Fridriksson; ¹University of South Carolina Department of Communication Sciences and Disorders, ²University of South Carolina McCausland Center for Brain Imaging

Introduction Speech entrainment (SE) facilitates fluent speech production for speakers with nonfluent aphasia1,2. It is hypothesized that nonfluent speech results from a functional disconnection of anterior and posterior cortical regions and that SE may improve coherence between these regions for successful entrainment.3 Transcranial alternating current stimulation (tACS) delivers low, periodically-alternating currents to improve functional connectivity between targeted regions through the amplification and entrainment of endogenous oscillations.4 Previous work suggests that in-phase tACS (alternating current with 0° relative phase difference) improves behavioral performance while anti-phase stimulation (180°) results in impaired behavioral performance due to impeded network synchronization. The purpose of the current study is to determine if tACS improves speech output in an SE task in patients with nonfluent aphasia. Methods 18 patients with chronic, nonfluent poststroke aphasia (mean age = 66 years; months post onset = 89) were enrolled in a double-blind, pseudorandomized study. 1mA of HD-tACS was delivered at 7Hz to residual anterior (IFGpo) and posterior (pMTG) regions of interest in the left hemisphere for 25 minutes across three different stimulation conditions: 1) in-phase; 2) anti-phase; and 3) sham while patients participated in SE. Individualized montages were created for each participant to estimate current flow and account for brain damage secondary to stroke (Soterix Medical, Inc.). Patients’ productions were recorded for transcription.5 Results We identified a significant main effect of condition for a secondary outcome measure, the number of different words produced: χ2(2) = 5.94, p = 0.05 where patients produced more words in the ‘in-phase’ condition as compared to ‘anti-phase’ and ‘sham’ conditions. Although not statistically significant, the primary behavioral outcome measure (proportion of correct words as compared to the script) and secondary measures such as the number of words per minute demonstrated a higher median for the ‘in-phase’ condition while number of errors reveals a higher median for the ‘anti-phase’ condition. Spectral-temporal analysis used mel-frequency cepstral coefficients in a dynamic time warping algorithm to examine the distance between the AV model and patient productions during the task. Results suggest that patients’ speech was better aligned (as evidenced by a smaller distance between the model and pt) during the stimulation conditions as compared to sham. Retrospective neurophysiological data suggest that patients who demonstrated better behavioral performance during the ‘in-phase’ stimulation, had greater preservation of the inferior temporal gyrus (z = 4.26) and poorer coherence as measured by rsfMRI and DTI between anterior and posterior regions (e.g. DTI: L insula to middle temporal gyrus; z = -4.24 and rsfMRI: inferior frontal gyrus to middle temporal gyrus). Conclusion/Summary Preliminary data suggest that tACS may not only improve speech output in speakers with nonfluent aphasia during a speech entrainment task, but also improve fluency, as evidenced by linguistic data (number of words per minute) and temporal data (mel-frequency cepstral coefficients). Outcomes from this proof-of-concept study encourage future investigations of tACS as an adjuvant for aphasia rehabilitation.

Global Motor Inhibition Precedes the Initiation of Stuttered Speech

Joan Orpella¹, Graham Flick¹, Florencia Assaneo², Eric S. Jackson¹; ¹New York University, ²Universidad Autónoma de México

Stuttering is a neurodevelopmental communication disorder that manifests itself, most saliently, as intermittent interruptions in speech. Despite progress towards discovering structural and functional abnormalities in the brains of stutterers compared to fluent speakers (i.e., trait differences), little is known about the neural bases of stuttered speech (i.e., state differences). This is because neuroimaging studies have focused on neural activity associated with fluent speech in stutterers, primarily due to the difficulty of reliably eliciting stuttered speech during laboratory testing. In this study, we simulated a real-life speaking situation in which the speaker knows the word they are about to say (e.g., their own name) and is then given a cue signaling the impending requirement to speak the word. To elicit a balanced number of stuttered and fluent speech, we leveraged a recently introduced clinical interview procedure in which participants identify anticipated words (words perceived as likely to be stuttered). In the motor literature, global motor inhibition is understood as an adaptive response to prevent the execution of motor actions. Global motor inhibition is typically observed as enhanced beta power in response to No-Go signals in regions of the action-stopping network, such as the right pre-supplementary motor area (R-preSMA) and the basal ganglia. It has also been proposed that global motor inhibition impedes the initiation and sequencing of speech motor commands in stuttering. Using magnetoencephalography (MEG), we tested the hypothesis that stuttered speech results from a global motor inhibition response. Twenty-nine adult stutterers participated. There were two experimental sessions: (1) Stuttering assessment and clinical interview following Jackson et al. (2021) to obtain participant-specific anticipated words likely to elicit stuttered speech. (2) MEG procedure during which participants produced 300 words (6 x 50 words, pseudorandomized) from the list constructed during the clinical interview. Words were visually presented, followed by a pre-cue and then a cue to speak. To determine differences in beta power between stuttered and fluent trials, we conducted a time-frequency decomposition separately for each type-trial (stuttered and fluent, equalized in counts). The decomposition was performed for frequencies in the beta band (12-30 Hz), and for times between the pre-cue and the cue to speak. To determine cortical origin, we conducted a power spectral density analysis in source space. Both fluent and stuttered trials showed the expected beta suppression pattern after the presentation of the pre-cue (i.e., when the participant becomes aware of the need to produce the anticipated word). However, stuttered trials were characterized by greater beta power (reduced beta suppression) compared to fluent trials. This power differential in the beta band originated in a single cluster corresponding to the R-preSMA. Results provide evidence that stuttered vs. fluent speech is associated with greater beta power in the R-preSMA, a node of the action-stopping network. This finding is in line with proposals that stuttered speech results from global motor inhibition, which could be the primary cause of stutterers' inability to initiate motor programs hypothesized in models of speech production.

Left posterior temporal cortex is the most critical brain region for recovery from aphasia

Sarah M. Schneck¹, Jillian L. Entrup¹, Caitlin F. Onuscheck¹, Deborah F. Levy¹, Dana K. Eriksson², Maysaa Rahman¹, L. Taylor Davis¹, Michael de Riesthal¹, Howard S. Kirshner¹, Stephen M. Wilson¹; ¹Vanderbilt University Medical Center, ²University of Arizona

In this cross-sectional study, we investigated neuroplasticity in post-stroke aphasia, making a concerted effort to ameliorate previously identified methodological concerns [1], and using multivariable models to disentangle the effects of structural damage and functional activation on behavior. Participants included 67 individuals with chronic or late sub-acute post-stroke aphasia and 46 neurotypical controls. Language function was measured with the Quick Aphasia Battery [QAB; 2] and varied across individuals with aphasia (QAB overall mean 7.5 ± 2.5 out of 10; range 0.7–9.9). Language regions were mapped using an adaptive semantic matching paradigm [3] that can reliably identify left-hemisphere language regions and minimizes effort confounds; both groups were able to perform the scanner task above chance. Whole brain analyses were corrected for multiple comparisons using permutation testing (voxelwise p < 0.005, corrected p < 0.05). Region of interest analyses used individually defined ROIs [4] to allow for the precise locations of language regions to differ across individuals. We found that language remained largely lateralized to the left hemisphere in aphasia. A direct comparison between groups revealed reduced activation in people with aphasia in several left-hemisphere language regions and the contralateral cerebellum. In the aphasia group, functional activation was positively correlated with language outcome (QAB overall score) in similar left-hemisphere language regions. Critically, at the whole-brain level, there was no evidence for differential recruitment of the right hemisphere, or correlations between right hemisphere activation and language outcome. We next defined four regions of interest (ROIs) in the bilateral inferior frontal gyrus (IFG) and bilateral posterior superior temporal sulcus (pSTS), and investigated relationships between structural damage, functional activation, and behavior. The greater sensitivity of this approach did reveal a modest association between right pSTS activity and language outcome in a simple correlational analysis (p = .033); stronger associations were observed in the left hemisphere regions that emerged from the whole brain analysis. We then used multivariate lesion-symptom mapping to derive a structural predictor of language outcome [5], and asked whether functional activity was associated with language outcome above and beyond expectations based on damage. We found that only in the left pSTS did functional activity remain a significant predictor of language outcome (p = .006). The right pSTS and the left IFG were no longer predictive of behavior. In sum, our data suggest that people with aphasia continue to process language in spared left-hemisphere language regions. Our whole-brain findings—that between-group differences are observed in left-hemisphere language regions, and that activation in left-hemisphere language regions is correlated with overall language function—directly mirror the two most compelling findings from our previous meta-analysis [1]. We found no compelling support for differential recruitment of right-hemisphere homotopic regions. Instead, our data provide strong evidence that left posterior temporal cortex remains the most critical neural substrate for recovery of language processing in post-stroke aphasia. References: [1] Wilson & Schneck, Neurobiol Lang 2021;2:22-82. [2] Wilson et al., PLoS One 2018;13(2):e0192773. [3] Wilson et al., Hum Brain Mapp 2018;39:3285-307. [4] Fedorenko et al., J Neurophysiol 2010;104:1177-1194. [5] Levy, Dissertation; 2021.

Premature brain aging is associated with aphasia severity mediated by compromised neural network controllability in the posterior superior temporal gyrus

Janina Wilmskoetter¹, Natalie Busby², Xiaosong He³, Lorenzo Caciagli⁴, Rebecca Roth⁵, Kathryn A. Davis⁴, Chris Rorden², Dani S. Bassett^4,6, Julius Fridriksson², Leonardo Bonilha⁵; ¹Medical University of South Carolina, ²University of South Carolina, ³University of Science and Technology of China, ⁴University of Pennsylvania, ⁵Emory University, ⁶Santa Fe Institute

Introduction: Aphasia recovery in the chronic stroke stages primarily depends on residual brain regions outside of the lesion. The integrity of the residual brain tissue can be directly measured by the tissue’s brain age, which is a significant predictor of aphasia severity independent of the characteristics of the lesion. We sought to investigate the underlying pathophysiology relating premature brain age (older biological brain age than chronological age) to aphasia severity in a cross-sectional study of 93 individuals with chronic aphasia. We hypothesized that the structural, dynamic embedding of language-specific regions within the remaining network, termed controllability (Gu et al., 2015; Wilmskoetter et al., 2021), mediates the relationship between premature brain age and aphasia severity. Additionally, we sought to validate the relationship between brain age and neural network controllability in an independent cohort of 54 age-matched individuals without a stroke. Methods: For each participant, brain-predicted age was estimated from the structural T1-weighted images using the freely available brainageR (v2.1) pipeline (Cole et al., 2017). We also calculated the brain age gap as the absolute difference between chronological and brain age. Higher (lower) values of the brain age gap reflect older (younger) brain age than chronological age. Further, we calculated modal controllability for a pre-selected set of core brain regions in the left hemisphere that are involved in language processing (Fedorenko et al., 2010). This set of regions spanned from the left frontal to the temporal and parietal areas. For the stroke group, we determined aphasia severity using the Aphasia Quotient of the Western Aphasia Battery (Revised; WAB-AQ) (Kertesz, 2007). Results: Brain age gap predicted WAB-AQ (β=-0.70, p=0.015) independent of chronological age and total lesion volume. Further, we observed a significant direct effect of brain age gap on WAB-AQ (effect=-0.74, standard error (SE)=0.29, 95% confidence interval (CI)=-1.32 to -0.16, p=0.013), as well as a significant indirect effect of brain age gap on WAB-AQ mediated by the modal controllability of the posterior STG (effect=0.13, bootstrapping SE=2.64, 95% CI=-1.19 to -0.04). Specifically, a higher brain age gap was associated with higher modal controllability of the posterior STG, and in turn, higher modal controllability of the posterior STG was associated with lower (worse) WAB-AQ scores. With a multivariable linear regression, we corroborated the significant relationship between the modal controllability of the posterior STG and the brain age gap in the non-stroke group after controlling for chronological age (β=-0.280, 2-tailed p=0.033). Conclusions: The effects of premature brain age on aphasia severity are significantly mediated by the modal controllability of the posterior superior temporal gyrus. Premature brain aging compromises widespread dynamic network mechanisms measured as modal controllability of language regions, in individuals with and without stroke, and predicts aphasia severity in individuals with chronic stroke.