Investigating the laminar profile of predictive signalling in human V1 with 7T fMRI

Poster No:

2181 

Submission Type:

Abstract Submission 

Authors:

Chantal Miller1, Alice Hickling1, Joost Haarsma1, Liliana Galindo1, Colleen Rollins1, Catarina Rua2, Christopher Rodgers2,3, Floris de Lange4, Peter Kok5, Jane Garrison1, Graham Murray1

Institutions:

1University of Cambridge, Cambridge, Cambridgeshire, UK, 2Wolfson Brain Imaging Centre, University of Cambridge, Cambridge, Cambridgeshire, UK, 3Oxford Centre for Clinical Magnetic Resonance Research, University of Oxford, Oxford, UK, 4Radboud University , Donders Institute for Brain, Cognition and Behaviour, Nijmegen, Gelderland, 5Wellcome Centre for Human Neuroimaging, University College London, London, UK

First Author:

Chantal Miller  
University of Cambridge
Cambridge, Cambridgeshire, UK

Co-Author(s):

Alice Hickling  
University of Cambridge
Cambridge, Cambridgeshire, UK
Joost Haarsma  
University of Cambridge
Cambridge, Cambridgeshire, UK
Liliana Galindo  
University of Cambridge
Cambridge, Cambridgeshire, UK
Colleen Rollins  
University of Cambridge
Cambridge, Cambridgeshire, UK
Catarina Rua  
Wolfson Brain Imaging Centre, University of Cambridge
Cambridge, Cambridgeshire, UK
Christopher Rodgers  
Wolfson Brain Imaging Centre, University of Cambridge|Oxford Centre for Clinical Magnetic Resonance Research, University of Oxford
Cambridge, Cambridgeshire, UK|Oxford, UK
Floris de Lange  
Radboud University , Donders Institute for Brain, Cognition and Behaviour
Nijmegen, Gelderland
Peter Kok  
Wellcome Centre for Human Neuroimaging, University College London
London, UK
Jane Garrison  
University of Cambridge
Cambridge, Cambridgeshire, UK
Graham Murray  
University of Cambridge
Cambridge, Cambridgeshire, UK

Introduction:

7T fMRI allows investigation of the cortical laminar profiles of feedforward and feedback (predictive) activity in humans. A 7T laminar fMRI study by Kok et al. (2016) used a Kanizsa illusion paradigm with 10 subjects, finding feedforward activity mainly in middle and superficial V1 layers and feedback activity in deep V1. We sought to replicate this result in a larger sample (n=22).

Methods:

22 healthy subjects were scanned in the 7T Terra MRI (Siemens) at the University of Cambridge, with anatomical (MP2RAGE, 0.65mm iso) and functional (GE-EPI, 0.8mm iso) sequences.
The task used the Kanizsa illusion (see Kok et al., 2016), consisting of pacman-shaped 'inducers' arranged to produce an illusory triangle (a feedback effect). Subjects completed 3 Kanizsa runs with Kanizsa trials (inducers arranged to produce the illusion), control trials (inducers arranged randomly) and fixation trials (only fixation point onscreen) and a flashing 'checkerboard' run. To maintain fixation, subjects did a visual search task in all trials.
Anatomy pre-processing was in FreeSurfer (recon-all), realignment in SPM. Co-registration of the anatomical volume to the mean functional volume involved volume-based, boundary-based and recursive boundary-based registrations. Our region of interest (ROI) consisted of V1 voxels with a receptive field centred on the illusion.
A 7T toolbox (https://github.com/TimVanMourik/OpenFmriAnalysis) defined 3 equivolume layers (deep, middle, superficial) and produced a design matrix (the layer distribution in each voxel in the ROI) and BOLD time courses for each layer and experimental run using a spatial GLM method (van Mourik et al., 2019). A temporal GLM regressed the layer time courses (for each layer and run) on stimulus presentation times to give beta parameter estimates for each layer and task condition. Parameter estimates were converted to percentage signal change by subtracting the fixation condition parameter from the parameter for each experimental (i.e., illusion, control or checkerboard) condition and dividing by the mean signal of the layer. Repeated-measures ANOVAs examined the effect of condition (feedback: Kanizsa v control, feedforward: checkerboard) on signal change across the 3 layers.

Results:

We present completed results for 10 subjects (analyses for additional subjects are ongoing).
We found no main effect of the illusion, F(1, 9) = .31, p = .593, η2 = .04 (no overall difference in activity caused by the illusion compared to control), and no interaction between trial condition and cortical depth, F(2, 18) = .18, p = .833, η2 = .02 (illusion did not significantly activate any layer; Figure 1). For the checkerboard, there was a significant effect of layer on the degree of signal change, F(2, 18) = 10.47, p < .001, η2 = .53, with a difference between the deep and superficial (p < .017) and the middle and superficial (p < .017) layers but no difference between the deep and middle layers (p = 0.200) (Figure 1).
Supporting Image: Figure1.jpg
 

Conclusions:

Feedforward activity (checkerboard run) produces greater activity in superficial than middle or deep V1 layers, similar to Kok et al. (2016). While Kok et al. (2016) found (illusion-evoked) feedback activity in deep V1, our results show the illusion produces no significant activity in any layer. We note that both this study and Kok et al. (2016) show considerable inter-subject variation in layer activity (Figure 2). A recent study found superficial but not deep layer activity associated with this illusion (Bergmann, Morgan & Muckli, 2019). Reasons for these conflicting results include possibilities of Type I and II errors; ROI selection is likely to strongly influence the results. Further, the Kanizsa illusion may not be represented in V1 but instead in higher-order visual areas (Murray et al, 2002a). Our study has implications for 7T methods, for example, by replicating the finding that the spatial GLM method reduces correlations between layer time courses, reducing partial volume effects (van Mourik et al., 2019).
Supporting Image: Figure2.jpg
 

Modeling and Analysis Methods:

Activation (eg. BOLD task-fMRI)

Neuroanatomy, Physiology, Metabolism and Neurotransmission:

Microcircuitry and Modules

Novel Imaging Acquisition Methods:

BOLD fMRI 2
Imaging Methods Other

Perception, Attention and Motor Behavior:

Perception: Visual 1

Keywords:

Cortical Layers
FUNCTIONAL MRI
HIGH FIELD MR
Perception
Vision

1|2Indicates the priority used for review

My abstract is being submitted as a Software Demonstration.

No

Please indicate below if your study was a "resting state" or "task-activation” study.

Task-activation

Healthy subjects only or patients (note that patient studies may also involve healthy subjects):

Healthy subjects

Was any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.

Yes

Was any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.

Not applicable

Please indicate which methods were used in your research:

Functional MRI

For human MRI, what field strength scanner do you use?

7T

Which processing packages did you use for your study?

SPM
FSL
Free Surfer

Provide references using author date format

Bergmann, J (2019), 'Two distinct feedback codes in V1 for ‘real’ and ‘imaginary’ internal experiences', BioRxiv (preprint)

Kok, P (2016), 'Selective Activation of the Deep Layers of the Human Primary Visual Cortex by Top-Down Feedback', Current Biology, vol. 26, no. 3, pp. 371-376

Murray, S (2002a), 'Shape perception reduces activity in human primary visual cortex', Proceedings of the National Academy of Sciences of the United States of America (PNAS), vol. 99, no. 23, pp. 15164-15169

van Mourik, T (2019), 'Laminar signal extraction over extended cortical areas by means of a spatial GLM', PLOS (Public Library of Science) ONE, vol. 14, no. 3, pp. 1-20