Electronic Thesis/Dissertation
 

Using Lesion-Symptom Mapping with Support Vector Regression (SVR-LSM) to Investigate the Neural Correlates of Inner Speech

Open Access Deposited

frontal lobe). The neural correlates of retrieval-based IS differ for the subjective and objective measures. Impairments on the objective measure of IS (picture-naming minus repetition) were associated with damage to structures supporting lexical-semantic retrieval (ventral stream

We found partial support for our hypotheses. SVR-LSM analyses revealed that the subjective measure of IS (IS self-report) relies on structures that support phonological storage and retrieval (ventral stream

inferior temporal lobe). For manipulation-based measures, impairments were associated with damage to structures supporting phonological storage and retrieval (ventral stream

Background

Data came from a preexisting dataset in which participants completed a large testing battery, including multiple measures of IS. Participants (n=67) were adult stroke survivors in the chronic phase of recovery. MRI scans were obtained to identify the precise location of lesions. Brain-behavior relationships were analyzed using lesion-symptom mapping with support vector regression (SVR-LSM). Hypotheses

Zero clusters passed the clusterwise threshold of statistical significance (p<0.05). Instead, the unthresholded voxelwise p-maps were analyzed to reveal underlying trends in brain-behavior relationships. Retrieval-based IS

superior temporal lobe) with diffuse patterns spreading towards sensorimotor regions (dorsal stream

Impairments on the subjective measure of IS (IS self-reports) were associated with damage to the superior temporal and frontal regions. Impairments on objective measures (picture-naming minus repetition) were associated with damage to the inferior temporal lobe. Manipulation-based IS

parietal lobe). Contrary to our hypothesis, the neural correlates supporting the subjective (IS self-report) and objective (picture-naming minus repetition) measures of retrieval-based IS showed minimal overlap. Results showed that damage impairing IS self-reports closely reflect the damage impairing manipulation-based measures of IS.

superior temporal lobe) and self-monitoring (dorsal stream

Impairments on rhyme judgment tasks were associated with damage to the superior temporal and parietal areas. Conclusions

silent retrieval of words (retrieval-based IS, e.g., self-report on silent picture naming) and silent retrieval plus manipulation of sounds (manipulation-based IS, e.g., silent rhyme judgment of two words).Questions

Impaired performance on retrieval-based IS will reflect damage to the ventral stream (e.g., temporal lobe). Impaired performance on manipulation-based IS will reflect damage to the ventral and dorsal streams (i.e., frontal/temporal lobes). Findings

This study focuses on inner speech (IS) in people with aphasia. Aphasia is a language disorder that occurs when neurological damage (e.g., stroke) disrupts the language network. A universal characteristic of aphasia is anomia, or difficulty naming. Many people with aphasia experience anomia despite self-reports of knowing the correct name. This phenomenon can be interpreted as an indication of successful IS, or the voice in one’s head. IS can be measured in multiple ways, capturing at least two definitions of IS

What are the neural correlates of IS? Do they differ for subjective and objective measures of retrieval-based IS? Do the neural correlates differ for retrieval- and manipulation-based definitions? Methods

Author Language Keyword Date created Type of Work License
  • All rights reserved
Rights statement GW Unit Degree Advisor Committee Member(s) Persistent URL

Notice to Authors

If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.

Thumbnail Title Date Uploaded Visibility Actions
Preview of Pfeffer_gwu_0075M_17017.pdf Pfeffer_gwu_0075M_17017.pdf 2025-07-20 Open Access