M1-PMd connectivity modulation via fMRI-neurofeedback
Poster No:
2065
Submission Type:
Abstract Submission
Authors:
Marine Keime1, Zeena-Britt Sanders2, Triin Ojakaar2, Cassandra Sampaio-Baptista1,2
Institutions:
1University of Glasgow, Glasgow, United Kingdom, 2University of Oxford, Oxford, United Kingdom
First Author:
Co-Author(s):
Cassandra Sampaio-Baptista
University of Glasgow|University of Oxford
Glasgow, United Kingdom|Oxford, United Kingdom
University of Glasgow|University of Oxford
Glasgow, United Kingdom|Oxford, United Kingdom
Introduction:
Research shows that brain connectivity during resting state highly corresponds to connectivity during a task and may predict individual difference in behavioural performance (Cheng et al., 2018). Neurofeedback (NF) could serve as a mean to explore the connection between resting state connectivity, task-related connectivity, and task performance.
Evidence shows greater M1-Pmd connectivity is associated with superior performance in action selection (AS) (Stewart, Tran & Cramer, 2014). However, the causal relationship has not been thoroughly examined.
Therefore, this study aimed to determine if M1-PMd connectivity could be modulated through covert fMRI-NF during rest, subsequently affecting cognitive-motor function.
Evidence shows greater M1-Pmd connectivity is associated with superior performance in action selection (AS) (Stewart, Tran & Cramer, 2014). However, the causal relationship has not been thoroughly examined.
Therefore, this study aimed to determine if M1-PMd connectivity could be modulated through covert fMRI-NF during rest, subsequently affecting cognitive-motor function.
Methods:
20 adults took part in this counterbalanced within-subject double-blind study. Participants were trained covertly with fMRI-NF in two separate conditions to increase and decrease M1-PMd connectivity.
The NF training was conducted in a 3T MRI scanner, consisting of approx 24 min (3 runs of 7min + rest). The feedback signal was fed back via a thermometer bar. The height represented M1-PMd correlation difference between the previous 20 TRs during NF and the M1-PMd correlation during the rest blocks. As a covert training, participants were just instructed that the higher the bar was, the more money they would earn.
The behavioural outcome of this study was measured by the action selection (AS) task, (O'Shea et al., 2007), which participants performed inside the MRI scanner before and after the NF.
The monetary incentive delay task (Knutson, Westdorp, Kaiser & Hommer, 2000) was used to measure reward sensitivity (RS), as previously shown to be associated with NF performance (Hellrung et al., 2019).
The NF training was conducted in a 3T MRI scanner, consisting of approx 24 min (3 runs of 7min + rest). The feedback signal was fed back via a thermometer bar. The height represented M1-PMd correlation difference between the previous 20 TRs during NF and the M1-PMd correlation during the rest blocks. As a covert training, participants were just instructed that the higher the bar was, the more money they would earn.
The behavioural outcome of this study was measured by the action selection (AS) task, (O'Shea et al., 2007), which participants performed inside the MRI scanner before and after the NF.
The monetary incentive delay task (Knutson, Westdorp, Kaiser & Hommer, 2000) was used to measure reward sensitivity (RS), as previously shown to be associated with NF performance (Hellrung et al., 2019).
Results:
During the NF training (vs rest), activity was present in various regions of the sensorimotor cortex (SMA, the precentral gyrus & PMd), see Figure 1B.
M1-PMd connectivity during the NF runs was analysed between conditions, using the extracted M1 and PMd timeseries. A repeated-measures ANOVA showed no main effect of condition (p=0.75), no main effect of run (p=0.42) and no condition*run interaction effect (F(2,36) = 0.40, p=0.68), see Figure 1A. An order effect was tested for. A mixed-methods ANOVA (between order, within condition) revealed a main effect of order (p=0.04), indicating that participants who started with the increase condition, overall decreased their connectivity more.
Effects of NF training on AS performance were then investigated. For the increase condition, there was no difference in reaction time (RT) between before and after NF (Wilcoxon paired signed-rank test, Z=-1.09, p=0.28). However, in the decrease condition, participants were significantly faster after the NF training compared to before (Wilcoxon paired signed-rank test, Z=-2.18, p=0.03, see Figure A.). When testing for interaction effects (condition*time), no significant difference was found in RT change between the conditions (paired t-test, p=0.47).
In terms of RS, a positive correlation between RS and the overall NF performance was found (Spearman correlation r=0.39, p=0.10, Figure 2A). This relationship was due to the significant correlation between RS and NF performance in the decrease condition only, with no relationship being found with performance in the increase condition (Spearman correlation decrease: r=-0.7, p=0.002; Spearman correlation increase: r=0.02, p=0.95, see Figure 2B).
M1-PMd connectivity during the NF runs was analysed between conditions, using the extracted M1 and PMd timeseries. A repeated-measures ANOVA showed no main effect of condition (p=0.75), no main effect of run (p=0.42) and no condition*run interaction effect (F(2,36) = 0.40, p=0.68), see Figure 1A. An order effect was tested for. A mixed-methods ANOVA (between order, within condition) revealed a main effect of order (p=0.04), indicating that participants who started with the increase condition, overall decreased their connectivity more.
Effects of NF training on AS performance were then investigated. For the increase condition, there was no difference in reaction time (RT) between before and after NF (Wilcoxon paired signed-rank test, Z=-1.09, p=0.28). However, in the decrease condition, participants were significantly faster after the NF training compared to before (Wilcoxon paired signed-rank test, Z=-2.18, p=0.03, see Figure A.). When testing for interaction effects (condition*time), no significant difference was found in RT change between the conditions (paired t-test, p=0.47).
In terms of RS, a positive correlation between RS and the overall NF performance was found (Spearman correlation r=0.39, p=0.10, Figure 2A). This relationship was due to the significant correlation between RS and NF performance in the decrease condition only, with no relationship being found with performance in the increase condition (Spearman correlation decrease: r=-0.7, p=0.002; Spearman correlation increase: r=0.02, p=0.95, see Figure 2B).
Conclusions:
Overall, participants did not manage to modulate their M1-PMd connectivity at rest using fMRI-NF, resulting in no changes in cognitive-motor function. The expected areas were however being modulated (sensorimotor system), indicating some level of successful implicit learning via the covert paradigm. This suggests more training could potentially lead to the desired effect. The link between reward sensitivity and NF performance gives strong evidence for a prediction tool of training success.
Motor Behavior:
Motor Planning and Execution 1
Novel Imaging Acquisition Methods:
BOLD fMRI 2
Keywords:
Cognition
FUNCTIONAL MRI
Motor
Other - Neurofeedback
1|2Indicates the priority used for review
·Figure 1
·Figure 2
Provide references using author date format
O'Shea, J., Johansen-Berg, H., Trief, D., Göbel, S., & Rushworth, M. F. (2007). Functionally specific reorganization in human premotor cortex. Neuron, 54(3), 479-490.
Stewart, J. C., Tran, X., & Cramer, S. C. (2014). Age-related variability in performance of a motor action selection task is related to differences in brain function and structure among older adults. Neuroimage, 86, 326-334.
Hellrung, L., Kirschner, M., Sulzer, J., Sladky, R., Scharnowski, F., Herdener, M., & Tobler, P. N. (2019). Individual differences in the mechnistic control of the dopaminergic midbrain. bioRxiv, (863639).
Knutson, B., Westdorp, A., Kaiser, E., & Hommer, D. (2000). FMRI visualization of brain activity during a monetary incentive delay task. Neuroimage, 12(1), 20-27.