Publication Date


Document Type


Committee Members

Frank Ciarallo (Committee Member), Subhashini Ganapathy (Advisor), Jason Parker (Committee Member), Jaime Ramirez-Vick (Committee Member), Curtis Tatsuoka (Committee Member)

Degree Name

Doctor of Philosophy (PhD)


There is a growing interest in the application of real-time (fMRI) with neurofeedback training (NFT) to the treatment of disorders associated with abnormal brain function. Chronic tinnitus is one such disorder that is often characterized by hyperactivity of the primary auditory cortex (A1) and decreased activity of the ventromedial prefrontal cortex (vmPFC). In this work, we investigated the use of fMRI-NFT to teach self-regulation of A1 using directed attention strategies. The overall objective of the proposed study is to determine the efficacy of fMRI-NFT for potential treatment of tinnitus. Healthy participants were separated into two groups: the experimental group received real feedback regarding activity in the A1 while control group was supplied sham feedback yoked from a random participant in the experimental group and matched for fMRI-NFT experience. Twenty-seven healthy volunteers with normal hearing (defined as no more than 1 frequency < -40 dB on a standard audiogram) underwent five fMRI-NFT sessions, each consisting of 1 auditory fMRI to functionally localize the A1, and 2 closed-loop neuromodulation runs using feedback from A1. FMRI data were acquired at 3T using 2D, single-shot echo planar imaging (EPI) during all three runs. The auditory fMRI was comprised of alternating blocks without and with auditory stimulation (continuous white noise delivered at 90 dB via in-ear headphones). During each closed-loop neuromodulation run, subjects completed alternating blocks identified as a "relax" period (i.e., watch the bar) or a "lower" (i.e., lower the bar). Auditory stimulation (same as for the auditory fMRI) was supplied during both sets of blocks. A1 activity was continuously presented using a simple bar plot during the closed-loop neuromodulation runs and updated with each EPI volume. A set of four simple directed attention strategies were suggested before each scan session to provide examples of brain control techniques to lower the bar, but the subjects were explicitly instructed to use any mental strategy they preferred. Average A1 deactivation was extracted from each closed-loop neuromodulation run and used to quantify the control over A1 (A1 control). Additionally, behavioral testing was completed outside of the MRI on sessions 1 and 5, and at a 2-week follow-up. This consisted of a subjective questionnaire to assess attentional control (attentional control scale; ACS) and two quantitative tests: the attention to emotion task (AE) and a vigilance variant of the continuous performance task (CPT-X). The ACS total was computed according to the associated literature. The AE task was assessed for the impact of emotion on attention by computing the percentage change between the average latency for emotional and neutral trials. A sensitivity index (d') was computed from the CPT-X using signal detection theory. A 2x5x2 (group by session and run) repeated measures ANOVA was performed on A1 control followed by post hoc pairwise comparisons to evaluate the session by group interaction. It was determined that A1 control improved with training, and that A1 control during sessions 5 and 2 was significantly higher than session 1 only for the experimental group. Behavior was assessed by conducting 2x2 and 2x3 (group by session) repeated measures ANOVAs on each test score to evaluate the effects of fMRI-NFT. Separate ANOVAs were conducted due to three participants that did not complete the follow-up behavioral assessment. The control group showed a markedly reduced average impact of emotion on attention. However, no other effects were observed. Additionally, the change in A1 deactivation and the impact of emotion on attention were negatively correlated. A neural assessment consisting of measures of brain activity in response to auditory stimulation, resting-state networks, and steady-st...

Page Count


Department or Program

Ph.D. in Engineering

Year Degree Awarded




Included in

Engineering Commons