Medical College of Wisconsin Cognitive Neuroscience Research Program

Sara Pillay, PhD, Principal Investigator

Abstract
Speech therapy treatments for post-stroke aphasia are of limited effectiveness, and patients are often left with residual deficits. There is a need for new inventions that have the potential to improve treatment outcomes. One such new approach is transcranial direct current stimulation (tDCS), a non-invasive brain stimulation method that uses a weak electrical current to modulate neuronal resting membrane potentials. Despite promising clinical trial results, it remains unclear how tDCS affects different language processes during recovery. The objectives of this study are to (1) study the behavioral effects of tDCS on language processes during recovery after stroke, and (2) examine how tDCS affects neural connectivity using resting-state fmri (rs- fMRI). Patients identified as having a phonologic (word-sound) impairment but intact semantics (word- meaning) will receive high-density array anodal or sham tDCS concurrent with either standard speech therapy or semantic feature analysis (SFA), a targeted speech therapy intervention aimed at recruiting the undamaged semantic system. Patients will be randomly assigned to receive standard speech therapy or SFA, and will receive either anodal- or sham-tDCS for 10 sessions (Treatment A) before crossing-over to receive a second 10 sessions (Treatment B) with whatever tDCS invention they did not receive initially. Patients and therapists will be blind to tDCS status. We propose to stimulate the left angular gyrus (AG), an area implicated in several language processes including semantic access. Patients will complete a language assessment plus rs-fMRI to at 3 time-points (pre-therapy, post-Treatment A, post-Treatment B). We have previously shown that BOLD activation in this region is associated specifically with successful responses during word reading in chronic aphasia patients with phonologic retrieval deficits but intact semantics. We hypothesize that the left AG has the potential to 'boost' phonological retrieval through activation of associated semantic information.

Lisa Conant, PhD, Principal Investigator

Abstract
Verbs play a fundamental role in the interface of semantics and syntax, and there is evidence that they may be particularly vulnerable to disruption in both developmental and acquired language disorders due to their semantic complexity. Verb meanings are bipartite in structure, with each verb defined by both idiosyncratic semantic content and a structural component shared across verbs of a specific class. However, the nature of their conceptual representation and its neural instantiation remain little understood. There is now significant evidence that concepts are learned through a process of generalization from real-world experiences and are subsequently partially 'embodied' in the modal systems involved in these experiences, with more abstract representation of conceptual knowledge in heteromodal 'hubs'. We previously developed a model of experiential attributes including not only sensory and motor domains but also affective, social, and other more abstract attributes, and have used it successfully to predict behavioral and brain data; however, this model does not adequately capture the structural component of verb meaning, known as event structure. In the proposed project, we plan to decompose event structure categories into meaningful dimensions based on how they are experienced, and incorporate these into our conceptual model. We will use representational similarity analysis applied to fMRI data collected during verb processing to identify brain regions sensitive to variations in these dimensions and to identify differences in the neural representation of aspectuality and argument structure. We hypothesize that the full model will be associated with activity in widely distributed regions comprising the semantic network, and that information about event structure will be associated with activity in left inferior frontal, posterior temporal, and inferior parietal regions. A better understanding of how verbs are represented could have clinical significance with regard to the development of more sensitive assessment tools and targeted interventions for language disorders.

Leonardo Fernandino, PhD, Principal Investigator

Abstract
Acquired language impairments resulting from brain injury are known as aphasias. According to the National Institutes of Health, 180,000 new cases of aphasia are diagnosed each year in the United States. The pattern of language impairments varies depending on the location and extent of the lesion, but many patients present with specifically semantic deficits. A major unexplained finding in aphasia research is that selective semantic deficits for animal concepts are substantially more common than selective deficits for human-made artifacts. The present project will address this issue by investigating the nature of the information encoded in the neural activity patterns underlying these concepts in healthy participants. Specifically, we will evaluate the extent to which different types of information (e.g., categorical, experiential, distributional) and different neurocognitive systems contribute to the representation of animal and artifact concepts, with the aim of elucidating the mechanisms that drive their differential rate of impairment. First, we will use functional MRI to evaluate the overall discriminability of the neural representations of animals and artifacts. We predict that, compared to animals, artifacts will be associated with more discriminable neural patterns. We will then use representational similarity analysis to characterize the information content of these patterns. We hypothesize that both animal and artifact concepts rely substantially more on experiential features than on other types of information. Finally, we will evaluate the relative contributions of different experiential features to the neural representation of these concepts. We expect that animal concepts will be associated with a more restricted set of experiential features, particularly visual features. These analyses will be conducted on the same data set, acquired from healthy participants while performing a semantic judgment task on individual nouns. The results will help advance our understanding of semantic language disorders and provide answers to a long-standing question in the field.

Laura Glass Umfleet, PhD, Principal Investigator; Yang Wang, MD, PhD, Co-PI

Abstract
Cerebrovascular injury can lead to executive dysfunction, which is a top predictor of decline in independent functioning. Reactivity of glial cells (crucial for CNS homeostasis) plays a critical role in cerebrovascular disease (CVD) seen in neurodegenerative conditions. Examining the neurovascular unit (NVU) provides an opportunity for understanding the intersection between cerebrovascular and glial pathology. NVU dysfunction and reductions in cerebral blood flow (CBF) can occur in dementia, and NVU dysfunction can precede cerebral atrophy. A promising biomarker of CVD is measurement of regional CBF responses, which are normally controlled by activity within the NVU, a process known as neurovascular coupling. Interrogation of this process will yield new insights into cerebrovascular correlates of cognition and neurodegeneration and improving neurovascular coupling could be neuroprotective. The main objectives of this proposal are to examine neurovascular coupling/uncoupling in prodromal dementia (mild cognitive impairment/MCI) and neurovascular coupling/uncoupling predictors of executive functioning. Executive dysfunction is common in non-amnestic MCI (naMCI), and, in particular, the dysexecutive naMCI phenotype is at greater risk for frontosubcortical network dysfunction. Establishing cerebrovascular biomarkers for naMCI phenotypes is crucial for determining outcome, etiological diagnosis/es, and neuroprotective treatments. In this pilot study, we will apply an innovative multiband and multiecho ASL/BOLD sequence to study neurovascular uncoupling in a naMCI cohort with executive dysfunction. This will allow us to achieve our first aim to evaluate MCI group differences in neurovascular coupling. We hypothesize that neurovascular uncoupling will emerge in mesial-temporal circuitry in aMCI and extra-temporal circuitry in naMCI. Our second aim is to examine possible physiological mechanisms that predict executive dysfunction in naMCI. We hypothesize that the neurovascular uncoupling in executive networks will be associated with fMRI executive functioning task performance. Our efforts will result in NIH funding proposals to elucidate mechanisms underlying cerebrovascular pathophysiology to aid in identifying therapeutic targets for neurodegenerative diseases.

Priyanka Shah-Basak, PhD, Principal Investigator

Abstract
The overarching objective is a systematic investigation of large-scale oscillatory synchronization underlying lexical-semantic and phonological language processing using magnetoencephalography (MEG) and transcranial alternating current stimulation (tACS). The long-term goal is to provide a critical foundation of evidence to support future studies examining the efficacy of tACS for the treatment of aphasia. Neuronal oscillations are ubiquitous in the human brain. Coherent or synchronous neuronal oscillations that emerge between brain areas are thought to facilitate configuration of functional networks to promote effective and selective transfer of information across regions in a task-dependent manner. Long-range synchronization (LRS) patterns between anterior and posterior perisylvian language regions, underlying component language processes, are poorly understood. Most prior studies have examined oscillatory correlates using spectral power analyses, which quantify local intra-regional synchronization but do not elucidate coupling between regions in the language network. The use of brain stimulation techniques such as tACS to examine the functional significance of LRS in language processing is a fledgling but promising research field. TACS involves delivering sinusoidal currents that can entrain endogenous neuronal oscillatory activity and influence LRS between stimulated regions. TACS thus provides a novel method for examining the physiology of the language network and the causal role of LRS in language processing, but little is known about how parameters such as phase (of alternating currents) affect network function. Additionally, prior tACS language studies have fallen short in their evaluations of LRS patterns in response to tACS phase conditions. To address these gaps, a double-blind, within-subject, sham-controlled experiment is proposed, using MEG to evaluate tACS effects on phase synchronization of oscillatory activity between anterior-posterior language areas during category-exemplar (semantic) and nonword rhyme matching (phonological) tasks, and effects of in-phase and anti-phase 10-Hz tACS on phonological processing and LRS. Our findings will provide a foundation for understanding dynamic functional coupling related to lexical-semantic and phonological processing in healthy language networks.