Healthy
Conditions
Keywords
functional magnetic resonance imaging, Transcranial Magnetic Stimulation, dorsolateral prefrontal cortex
Brief summary
The purpose of this study is to improve understanding of the way transcranial magnetic stimulation (TMS), a form of non-invasive brain stimulation, affects the brain. The study hypothesis that when theta burst stimulation (TBS) is applied during a controlled mental state, network changes will be facilitated, compared to stimulation when mental state is uncontrolled. This study will focus on the dorsolateral prefrontal cortex (dlPFC) and the associated frontoparietal network (FPN), which subserves cognitive control - the ability to flexibly adapt and regulate behavior, an ability known to be impaired in neuropsychiatric conditions such as depression and dementia. Healthy volunteers that qualify for this study will have psychological assessments and cognitive measures (due to Covid, some of these were done via teleconference), as well as functional Magnetic Resonance Imaging (fMRI) scans, completed after administration of TMS. Participants will be asked to come in for a total of five visits that include; a screening and assessment visit; a baseline functional magnetic resonance imaging (fMRI) scan, followed by TMS session; Visits 3, 4, and 5 will be the experimental TMS session, followed by fMRI scan.
Detailed description
We will test the broad hypothesis that when TBS is applied during a controlled mental state, network changes will be facilitated, compared to stimulation when mental state is uncontrolled. We will focus on the dorsolateral prefrontal cortex (dlPFC) and the associated fronto-parietal network (FPN), which subserves cognitive control -- the ability to flexibly adapt and regulate behavior, an ability known to be impaired in neuropsychiatric conditions such as depression and dementia. We will use an 'n-back' task tapping cognitive control and the FPN. We will employ a within-subjects design with 40 healthy subjects in 4 MRI sessions. Each MRI session will consist of blood oxygenation level-dependent (BOLD) fMRI during an n-back task, resting state BOLD fMRI to measure connectivity and resting state arterial spin labeling (ASL) MRI to measure cerebral blood flow (rCBF) and examine effects on resting activity level. BOLD activation during the n-back will identify the FPN and the target site for dlPFC TBS. After a baseline fMRI session, subsequent sessions over different days will entail TBS, immediately followed by an MRI session to assess the effects of stimulation. TBS will involve: 1) dlPFC stimulation by active iTBS (600 pulses) alone or 2) while simultaneously performing an n-back cognitive task or 3) vertex (control) iTBS stimulation, alone.
Interventions
DEVICETMS
Intermittent theta burst stimulation TMS applied to the cortex to excite cerebral cortex
BEHAVIORALn-back working memory task
Subjects perform an executive function task, in which they view the serial presentation of letters and decide whether or not a letter matches a letter presented 'n' letters back (2 letters or 1 letter)
Sponsors
National Institute of Mental Health (NIMH)
University of Michigan
Study design
Allocation
NON_RANDOMIZED
Intervention model
CROSSOVER
Primary purpose
BASIC_SCIENCE
Masking
SINGLE (Subject)
Masking description
All subjects will receive all interventions in a cross-over design. The three intervention sessions (visits 3, 4 & 5) will be given in counter-balanced order, stratified by gender. Subjects will be blind to the nature of the questions being asked; however, they will be told that different stimulation paradigms will be used, which will be evident to the subjects as they go through the procedures.
Intervention model description
All participants will have 3 MRI sessions subsequent to the baseline visit and these will occur in a counterbalanced order.
Eligibility
Sex/Gender
ALL
Age
18 Years to 50 Years
Healthy volunteers
Yes
Inclusion criteria
* Women of child bearing age can not be pregnant or trying to become pregnant * Ability to tolerate small, enclosed spaces without anxiety * Size compatible with scanner gantry, e. g. men over 6 feet tall that weigh more than 250 lbs, men under 6 feet tall that weigh over 220 lbs, women over 5'11 tall that weigh more than 220 lbs, or women under 5'10 tall that weigh more than 200 lbs. Subjects of these weights or greater typically have difficult fitting into the fMRI scanner properly * Ability and willingness to give informed consent to participate * Alcohol or drug dependence (if in remission for greater than 5 years)
Exclusion criteria
* History of past or current mental illness (except simple phobias) * History of closed head injury, for example, loss of consciousness \> approximately 5 minutes, hospitalization, neurological sequela; * Metals, implants or metallic substances within or on the body that might cause adverse effects to the subject in a strong magnetic field, or interfere with image acquisition (for example; aneurysm clips, retained particles or metal workers with exposures, neurostimulators, foil-backed transdermal patches, carotid or cerebral stents, cerebral spinal fluid (CSF) shunts; magnetic dental implants, ferromagnetic ocular implants, pacemakers, and automatic implantable defibrillators). * Prescription or non-prescription, with psychotropic effects (birth control medications allowed) * First-degree family members with a history of epilepsy * History of serious neurological illness or current medical condition that could compromise brain function, such as liver failure
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Accuracy to 2-back | 60 minutes after TMS during fMRI | Correct responses to letter stimuli, as a percentage of all responses |
| 2-back Minus 1-back Blood Oxygen Level-Dependent (BOLD) Activation, Voxelwise in FPN | 60 minutes after TMS during fMRI | Fronto-parietal network (FPN) defined by BOLD change while subject performed the n-back working memory task, contrasting high (2-back) versus low (1-back) loads. Using the SPM12 package, data were normalized per standard, open source routines. Using a General Linear Model framework, a model was estimated with regressors (after convolution with hemodynamic response function) for 2-back & 1-back conditions for each subject to predict BOLD change each day (arm).For analysis of group effects, second-level, between-subject analyses on normalized images of the 2-back minus 1-back beta estimate from the first level were entered into regression models, with mean frame displacement as a co-variate of no-interest to test contrasts between the arms/interventions. Using a FPN mask, the eigenvalues from the second-level estimates were extracted and entered into the analysis as an outcome measure. Note: eigenvalues are arbitrary units. Larger values indicate more BOLD signal. |
| Frontoparietal Network (FPN) Connectivity to Dorsolateral Prefrontal Cortex (dlPFC) Theta Burst Stimulation (TBS) Target | 60 minutes after TMS during fMRI | Analysis of resting-state connectivity was performed used the CONN toolbox, using standard techniques to demonstrate connectivity between a spherical seed placed on each participant's locus of dlPFC stimulation, and the rest of the brain. Connectivity (correlations of BOLD signal) was first calculated for each participant, and then spatially averaged in MNI brain space, between participants. A 'cluster' of connectivity was identified, only if the number of voxels (thresholded at P \<0.001) exceeded the count of 25. The outcome measure here is a count of the number of clusters exceeding this threshold, across all subjects. It represents significant connectivity between the site of stimulation and that cluster in the brain, for all subjects. |
| Cerebral Blood Flow (rCBF) at Stimulation Target | 15 minutes after TMS during fMRI | Regional cerebral blood flow measured at the site of theta burst stimulation (TBS) in milliliters per 100 mg tissue per minute |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| 2-back Minus 1-back BOLD Activation, Voxelwise in Whole Brain | 60 minutes after TMS during fMRI | Fronto-parietal network (FPN) defined by BOLD change while subject performed the n-back working memory task, contrasting high (2-back) versus low (1-back) loads. Using the SPM12 package, data were normalized per standard, open source routines. Using a General Linear Model framework, a model was estimated with regressors (after convolution with hemodynamic response function) for 2-back & 1-back conditions for each subject to predict BOLD change each day (arm).For analysis of group effects, second-level, between-subject analyses on normalized images of the 2-back minus 1-back beta estimate from the first level were entered into regression models. With a cluster threshold of voxel magnitude \< 0.001, clusters of difference are defined. Because there are often multiple clusters in a contrast, the number below is the size, in voxels, of the largest cluster, across all participants, for each session. |
| Measure Cerebral Blood Flow (rCBF) in FPN | 15 minutes after TMS during fMRI | Regional cerebral blood flow measured in the FPN in milliliters per 100 mg tissue per minute |
| Median Reaction Time (RT) in 2-back | 60 minutes after TMS during fMRI | Median reaction time for subjects responding in the n-back task, for correct responses (1-back and 2-back) |
| D-prime in 2-back | 60 minutes after TMS during fMRI | d-prime = z(H) - z(F) , where z(H) and z(F) are the z transforms of hit rate and false alarm, respectively. Hit rate = number of correctly identified targets/number of targets presented False alarm = number of incorrectly identified targets/number of non-targets presented |
Countries
United States
Participant flow
Pre-assignment details
Of 59 participants who were screened, 6 were ineligible and 53 were enrolled. With each treatment, regardless of the sequence, an MRI was performed.
Participants by arm
| Arm | Count |
|---|---|
| All Groups Since participants' data are only analyzed if they completed all assignments, and the order of assignment was immaterial to the trial, as it was a basic experiment studying humans, there was no reason to separate out the participants based on the order of interventions. Data is provided for all subjects whose data is analyzed. | 40 |
| Total | 40 |
Withdrawals & dropouts
| Period | Reason | FG000 |
|---|---|---|
| Overall Study | Adverse Event | 1 |
| Overall Study | COVID based withdrawal | 5 |
| Overall Study | No Motor Threshold obtainable | 1 |
| Overall Study | Physician Decision | 1 |
| Overall Study | Withdrawal by Subject | 5 |
Baseline characteristics
| Characteristic | All Groups |
|---|---|
| Age, Continuous | 25.5 years STANDARD_DEVIATION 7.4 |
| Ethnicity (NIH/OMB) Hispanic or Latino | 3 Participants |
| Ethnicity (NIH/OMB) Not Hispanic or Latino | 37 Participants |
| Ethnicity (NIH/OMB) Unknown or Not Reported | 0 Participants |
| Race (NIH/OMB) American Indian or Alaska Native | 0 Participants |
| Race (NIH/OMB) Asian | 8 Participants |
| Race (NIH/OMB) Black or African American | 5 Participants |
| Race (NIH/OMB) More than one race | 3 Participants |
| Race (NIH/OMB) Native Hawaiian or Other Pacific Islander | 0 Participants |
| Race (NIH/OMB) Unknown or Not Reported | 0 Participants |
| Race (NIH/OMB) White | 24 Participants |
| Region of Enrollment United States | 40 Participants |
| Sex: Female, Male Female | 27 Participants |
| Sex: Female, Male Male | 13 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk | EG002 affected / at risk | EG003 affected / at risk |
|---|---|---|---|---|
| deaths Total, all-cause mortality | 0 / 47 | 0 / 41 | 0 / 43 | 0 / 40 |
| other Total, other adverse events | 4 / 47 | 2 / 41 | 0 / 43 | 1 / 40 |
| serious Total, serious adverse events | 0 / 47 | 0 / 41 | 0 / 43 | 0 / 40 |
Outcome results
Primary
2-back Minus 1-back Blood Oxygen Level-Dependent (BOLD) Activation, Voxelwise in FPN
Fronto-parietal network (FPN) defined by BOLD change while subject performed the n-back working memory task, contrasting high (2-back) versus low (1-back) loads. Using the SPM12 package, data were normalized per standard, open source routines. Using a General Linear Model framework, a model was estimated with regressors (after convolution with hemodynamic response function) for 2-back & 1-back conditions for each subject to predict BOLD change each day (arm).For analysis of group effects, second-level, between-subject analyses on normalized images of the 2-back minus 1-back beta estimate from the first level were entered into regression models, with mean frame displacement as a co-variate of no-interest to test contrasts between the arms/interventions. Using a FPN mask, the eigenvalues from the second-level estimates were extracted and entered into the analysis as an outcome measure. Note: eigenvalues are arbitrary units. Larger values indicate more BOLD signal.
Time frame: 60 minutes after TMS during fMRI
Population: Of the 40 participants, image data from 3 failed quality control.
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| TMS to dlPFC, Without a Concurrent Task (Passive) | 2-back Minus 1-back Blood Oxygen Level-Dependent (BOLD) Activation, Voxelwise in FPN | 0.74 arbitrary units | Standard Deviation 0.54 |
| TMS to Vertex, Without Concurrent Task (Control) | 2-back Minus 1-back Blood Oxygen Level-Dependent (BOLD) Activation, Voxelwise in FPN | 0.75 arbitrary units | Standard Deviation 0.45 |
| TMS to dlPFC, During Task (Active) | 2-back Minus 1-back Blood Oxygen Level-Dependent (BOLD) Activation, Voxelwise in FPN | 0.58 arbitrary units | Standard Deviation 0.39 |
Comparison: Analysis used standard, open-source methods for pre-processing . First-level analysis used the general framework of the modified General Linear Model, implemented in SPM12 with temporal convolution. For the n-back activation task, the first-level design matrix included regressors for 2-back and 1-back conditions, as well as 24 movement parameters. Statistical testing was done to establish whether more or less BOLD (neural activity) change occurred between two conditions.p-value: =0.1195% CI: [-0.037, 0.355]t-test, 2 sided
Comparison: Analysis used standard, open-source routes for slice timing correction, field map correction, realignment, smoothing (6 mm Gaussian kernel) and spatial normalization (MNI-152). First-level analysis used the general framework of the modified General Linear Model, implemented in SPM12 with temporal convolution. For the n-back activation task, the first-level design matrix included regressors for 2-back and 1-back conditions, as well as 24 movement parameters.p-value: 0.9495% CI: [-0.245, 0.227]t-test, 2 sided
Primary
Accuracy to 2-back
Correct responses to letter stimuli, as a percentage of all responses
Time frame: 60 minutes after TMS during fMRI
Population: Due to technical failures, data was lost for one participant during control and for 24 during Active.
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| TMS to dlPFC, Without a Concurrent Task (Passive) | Accuracy to 2-back | 96.7 percentage correct | Standard Deviation 3.9 |
| TMS to Vertex, Without Concurrent Task (Control) | Accuracy to 2-back | 96.7 percentage correct | Standard Deviation 3.6 |
| TMS to dlPFC, During Task (Active) | Accuracy to 2-back | 96.2 percentage correct | Standard Deviation 4.5 |
p-value: 0.55ANOVA
p-value: 0.65ANOVA
Primary
Cerebral Blood Flow (rCBF) at Stimulation Target
Regional cerebral blood flow measured at the site of theta burst stimulation (TBS) in milliliters per 100 mg tissue per minute
Time frame: 15 minutes after TMS during fMRI
Population: pre-processing failures in 5 participants
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| TMS to dlPFC, Without a Concurrent Task (Passive) | Cerebral Blood Flow (rCBF) at Stimulation Target | 57.2 ml/100 mg tissue/min | Standard Deviation 20.7 |
| TMS to Vertex, Without Concurrent Task (Control) | Cerebral Blood Flow (rCBF) at Stimulation Target | 57.9 ml/100 mg tissue/min | Standard Deviation 28.5 |
| TMS to dlPFC, During Task (Active) | Cerebral Blood Flow (rCBF) at Stimulation Target | 58.4 ml/100 mg tissue/min | Standard Deviation 20.3 |
p-value: 0.91ANOVA
Primary
Frontoparietal Network (FPN) Connectivity to Dorsolateral Prefrontal Cortex (dlPFC) Theta Burst Stimulation (TBS) Target
Analysis of resting-state connectivity was performed used the CONN toolbox, using standard techniques to demonstrate connectivity between a spherical seed placed on each participant's locus of dlPFC stimulation, and the rest of the brain. Connectivity (correlations of BOLD signal) was first calculated for each participant, and then spatially averaged in MNI brain space, between participants. A 'cluster' of connectivity was identified, only if the number of voxels (thresholded at P \<0.001) exceeded the count of 25. The outcome measure here is a count of the number of clusters exceeding this threshold, across all subjects. It represents significant connectivity between the site of stimulation and that cluster in the brain, for all subjects.
Time frame: 60 minutes after TMS during fMRI
Population: 3 subjects failed image quality control processing
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| TMS to dlPFC, Without a Concurrent Task (Passive) | Frontoparietal Network (FPN) Connectivity to Dorsolateral Prefrontal Cortex (dlPFC) Theta Burst Stimulation (TBS) Target | 0 cluster count for connectivity |
| TMS to Vertex, Without Concurrent Task (Control) | Frontoparietal Network (FPN) Connectivity to Dorsolateral Prefrontal Cortex (dlPFC) Theta Burst Stimulation (TBS) Target | 10 cluster count for connectivity |
| TMS to dlPFC, During Task (Active) | Frontoparietal Network (FPN) Connectivity to Dorsolateral Prefrontal Cortex (dlPFC) Theta Burst Stimulation (TBS) Target | 0 cluster count for connectivity |
Comparison: The contrast reflects the difference in connectivity between the conditions/sessions, at the level of each subject, then spatially averaged across all subjects. Functional connectivity strength for the dlPFC seed was represented by Fisher-transformed bivariate correlation coefficients from a weighted general linear model (weighted-GLM), defined separately for each pair of seed and target areas, modeling the association between their BOLD signal time series.p-value: 1random field theory
Comparison: Results were thresholded using a combination of a cluster-forming p \< 0.001 voxel-level threshold, and a corrected false discovery rate (FDR) p \< 0.05 cluster-size threshold.p-value: 1random field theory
Secondary
2-back Minus 1-back BOLD Activation, Voxelwise in Whole Brain
Fronto-parietal network (FPN) defined by BOLD change while subject performed the n-back working memory task, contrasting high (2-back) versus low (1-back) loads. Using the SPM12 package, data were normalized per standard, open source routines. Using a General Linear Model framework, a model was estimated with regressors (after convolution with hemodynamic response function) for 2-back & 1-back conditions for each subject to predict BOLD change each day (arm).For analysis of group effects, second-level, between-subject analyses on normalized images of the 2-back minus 1-back beta estimate from the first level were entered into regression models. With a cluster threshold of voxel magnitude \< 0.001, clusters of difference are defined. Because there are often multiple clusters in a contrast, the number below is the size, in voxels, of the largest cluster, across all participants, for each session.
Time frame: 60 minutes after TMS during fMRI
Population: 3 participants were not analyzed due to poor data quality
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| TMS to dlPFC, Without a Concurrent Task (Passive) | 2-back Minus 1-back BOLD Activation, Voxelwise in Whole Brain | 1139 voxel count |
| TMS to Vertex, Without Concurrent Task (Control) | 2-back Minus 1-back BOLD Activation, Voxelwise in Whole Brain | 1532 voxel count |
| TMS to dlPFC, During Task (Active) | 2-back Minus 1-back BOLD Activation, Voxelwise in Whole Brain | 907 voxel count |
Comparison: Statistical thresholding using a random field model adjusted for multiple comparisons, with initial voxel magnitude/height threshold set at p \< 0.001 and peak and cluster-corrected significance levels obtained (false discovery rate \[FDR\]corrected), across the entire brain.p-value: 0.54random field theory
Comparison: Statistical thresholding using a random field model adjusted for multiple comparisons, with initial voxel magnitude/height threshold set at p \< 0.001 and peak and cluster-corrected significance levels obtained (false discovery rate \[FDR\]corrected), across the entire brain..p-value: 0.93random field theory
Secondary
D-prime in 2-back
d-prime = z(H) - z(F) , where z(H) and z(F) are the z transforms of hit rate and false alarm, respectively. Hit rate = number of correctly identified targets/number of targets presented False alarm = number of incorrectly identified targets/number of non-targets presented
Time frame: 60 minutes after TMS during fMRI
Population: Due to technical failures, data was lost for one participant during control and for 24 participants during Active.
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| TMS to dlPFC, Without a Concurrent Task (Passive) | D-prime in 2-back | 3.8 d-prime | Standard Deviation 0.7 |
| TMS to Vertex, Without Concurrent Task (Control) | D-prime in 2-back | 3.8 d-prime | Standard Deviation 0.6 |
| TMS to dlPFC, During Task (Active) | D-prime in 2-back | 3.7 d-prime | Standard Deviation 0.8 |
p-value: 0.41ANOVA
p-value: 0.5ANOVA
Secondary
Measure Cerebral Blood Flow (rCBF) in FPN
Regional cerebral blood flow measured in the FPN in milliliters per 100 mg tissue per minute
Time frame: 15 minutes after TMS during fMRI
Population: 5 subjects with poor data quality had to be excluded
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| TMS to dlPFC, Without a Concurrent Task (Passive) | Measure Cerebral Blood Flow (rCBF) in FPN | 52.7 mL/100 mg tissue/min | Standard Deviation 11.8 |
| TMS to Vertex, Without Concurrent Task (Control) | Measure Cerebral Blood Flow (rCBF) in FPN | 52.9 mL/100 mg tissue/min | Standard Deviation 15.6 |
| TMS to dlPFC, During Task (Active) | Measure Cerebral Blood Flow (rCBF) in FPN | 51.9 mL/100 mg tissue/min | Standard Deviation 13.1 |
Comparison: Repeated measures ANOVA to test for differences in cerebral blood flow (perfusion) in the frontoparietal networkp-value: 0.8ANOVA
Secondary
Median Reaction Time (RT) in 2-back
Median reaction time for subjects responding in the n-back task, for correct responses (1-back and 2-back)
Time frame: 60 minutes after TMS during fMRI
Population: Due to technical failures, data was lost for one participant during control and for 24 participants during Active.
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| TMS to dlPFC, Without a Concurrent Task (Passive) | Median Reaction Time (RT) in 2-back | 573.2 milliseconds | Standard Deviation 175.8 |
| TMS to Vertex, Without Concurrent Task (Control) | Median Reaction Time (RT) in 2-back | 556.4 milliseconds | Standard Deviation 164.2 |
| TMS to dlPFC, During Task (Active) | Median Reaction Time (RT) in 2-back | 572.1 milliseconds | Standard Deviation 135.4 |
p-value: 0.06ANOVA
p-value: 0.68ANOVA