Poster No:
928
Submission Type:
Abstract Submission
Authors:
Ashley Zhou1, Daniel Mitchell2, John Duncan3
Institutions:
1University of Cambridge, Cambridge, Cambridgeshire, 2MRC Cognition and Brain Sciences Unit, University of Cambridge, Cambridge, Cambridgeshire, 3MRC Cognition and Brain Sciences Unit, Cambridge, Cambridgeshire
First Author:
Ashley Zhou
University of Cambridge
Cambridge, Cambridgeshire
Co-Author(s):
Daniel Mitchell
MRC Cognition and Brain Sciences Unit, University of Cambridge
Cambridge, Cambridgeshire
John Duncan
MRC Cognition and Brain Sciences Unit
Cambridge, Cambridgeshire
Introduction:
Recent findings challenge views of the DMN as a purely task-negative or self-oriented network, by showing increased DMN activity during demanding external task-switches between cognitive domains, versus within-domain switches or task repeats (Crittenden et al., 2015; Smith et al., 2018). In this fMRI study, we examine how the DMN's response to task switches depends on current task set complexity, task expectancy, and instructional order. Results showed that while DMN activation at task switches was unaffected by either the number of currently relevant tasks, or task expectancy, it depended on the order in which groups of tasks had been learnt. This suggests sensitivity to an intrinsic task-set hierarchy, and a role in complex cognitive control processes. Results may also explain inconsistent observations of DMN activation across task-switch studies.
Methods:
36 healthy participants (13 male), aged 18-45, were tested. For each of four task domains (lexical, semantic, faces, shapes), two different tasks were cued by colored frames. Participants learned task-color pairs in two groups of two domains, and performed them in the scanner in the same groups (four 2-domain runs) or all intermixed (two 4-domain runs). Per trial, a colored frame cued a binary judgement on a central stimulus. The task sequence defined equal numbers of different transition types (e.g. task repeat, within-domain switch, between-domain switch). Task expectancy (variable by switch type, balanced per switch type) was varied across runs, crossed with domain number.
Data were acquired on a 3T Siemens Prisma MRI scanner, using T2*-weighted EPI (TR 1.2 s, TE 30 ms, flip angle 67°, 3×3x3 mm voxels, multiband factor 2). Pre-processing (using SPM12 and AutomaticAnalysis; Cusack et al., 2015 ) applied spatial realignment, slice-time correction, co-registration, and normalization to the MNI template, with no spatial smoothing.
A GLM was created by convolving the response periods of all trials per condition with the canonical HRF. Conditions counterbalanced each combination of switch type (task repeat, within-domain switch, between-domain switch, between-group-switch, restart, rest), number of active task domains (2 or 4), task expectancy (variable, balanced), and task group (learnt 1st or 2nd). Movement parameters and block means were added as covariates. Analysis focused on mean signal from a Core DMN ROI.
Results:
Mean reaction time (RT) was 1.1 s. Expected switch costs were seen, with RT of switch trials (mean across switch types) slower than for task repeats (t35 =10.1, p<0.01, BF>2x10^10).
Activity in Core DMN was significantly higher for averaged switch trials compared to task repeats (t35 =6.47, p<0.01, BF>4x10^5). Two-way ANOVA with factors of switch type (within-domain, between-domain) and number of domains (2, 4) showed a significant effect of switch type (F(1,35) =19.6, p<0.01, BF=270) but no effect of domain number (F(1,35)=1.15, p=0.29, BF=0.36) or interaction (F(1,35)=1.08, p=0.31, BF=0.35). Similarly, a two-way ANOVA with switch type and task expectancy as factors found no effect of task expectancy (F(1,35)=1.16, p=0.29, BF=0.36) or interaction (F(1,35) =0.65, p=0.43, BF=0.29). However, a two-way ANOVA with factors of switch type and instructed order found a significant interaction, with higher activity for between-domain switches of later-learned domains (F(1,35)=5.42, p=0.03, BF=2.17), but no main effect of instructional order (F(1,35)=0.02, p=0.89, BF=0.22).
Conclusions:
We investigated influences on DMN activation at task switches. Results suggest that DMN activity is sensitive to task structure complexity, but depends on learning order rather than the number of currently relevant tasks or particular task expectancies. As later-learned tasks increase complexity of the task set, DMN shows a different response profile across task transitions. Results speculatively suggest that a capacity-limited, two-level, hierarchical task model underlies DMN involvement in task transitions.
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making 1
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI) 2
Keywords:
Cognition
FUNCTIONAL MRI
1|2Indicates the priority used for review
Provide references using author date format
Cusack, R., Vicente-Grabovetsky, A., Mitchell, D. J., Wild, C. J., Auer, T., Linke, A. C., & Peelle, J. E. (2015). Automatic analysis (aa): Efficient neuroimaging workflows and parallel processing using Matlab and XML. Frontiers in Neuroinformatics, 8. https://www.frontiersin.org/articles/10.3389/fninf.2014.00090
Crittenden, B. M., Mitchell, D. J., & Duncan, J. (2015). Recruitment of the default mode network during a demanding act of executive control. eLife, 4, e06481–e06481. https://doi.org/10.7554/eLife.06481
Smith, V., Mitchell, D. J., & Duncan, J. (2018). Role of the Default Mode Network in Cognitive Transitions. Cerebral Cortex (New York, N.Y. 1991), 28(10), 3685–3696. https://doi.org/10.1093/cercor/bhy167