Synopsis

Cortical thickness changes have been shown to be correlated with a wide variety of behavioral measures. Until recently, methods to probe the laminar changes underlying these large scale cortical changes in vivo have not been available. Here, we present methods to measure laminar thickness within the cortex, and show that sufficient precision exists to observe behavioral correlates within individual layers. Furthermore, our data are consistent with the hypothesis that behavior learned early in life has different laminar thickness correlates than does behavior learned later in life.

Introduction

It is well established that cortical thickness can change as a function of experience, but less well understood how underlying cortical layers change to support this. Here, we attempt to measure differences in laminar thickness within the cortex of the brain and its relationship to measured behavior, with the hypothesis that different layers contribute to changes in cortical thickness depending on the age of the subject when the behavior was learned.

We focus on the middle fusiform face area (FFA2), where cortical thickness of the FFA correlates to face and vehicle recognition in opposite fashions^1,2, which we postulate may be due to differences in age and plasticity when each type of recognition expertise was learned. If so, we may expect that variability in cortical thickness associated with face recognition, learned early in life, could be specifically localized to the deep layers of cortex. In contrast, vehicle recognition expertise, learned later in life, may be better explained by variance in thickness of all layers, potentially due to differences in cortical plasticity and remodeling mechanisms^3-5.

Methods

(Subjects) 13 adult men aged 19 to 29 (mean age = 22.46+/-3.60yrs), were recruited to vary in face recognition ability.

(Behavioral Testing) In all, four behavioral tests were performed, including the extendedCambridge Face Memory Test6(CFMT+), the Vanderbilt Expertise Test7(VET),the Vanderbilt Face Matching Test (VFMT), and a sequential matching test used to quantify individual skill at matching cars, planes, birds and butterflies8. From these, an aggregate Face index was calculated based on standardized face performance on the CFMT+, VFMT, VET-Male, and VET-Female tests, as was an aggregate Non-Living Object index, and each was residualized relative to the other categories.

(Imaging) All subjects were imaged using a whole body 7T MRI scanner in combination with a quadrature, head only transmit coil and a 32 channel receive coil array. In each subject, imaging was separated into three stages consisting of whole brain anatomic imaging, functional localization, and ultrahigh resolution susceptibility weighted imaging.

Anatomic imaging consisted of 1mm isotropic MPRAGE images used for planning of all future scans. Functional localization of the FFA was performed using Faces vs Objects N-Back task, and was processed online to inform planning of high resolution susceptibility weighted imaging over the FFA. Six high resolution T2*weighted images were acquired as separate and independent acquisitions in each subject using slice selective cartesian gradient echo acquisitions (FOV=240x180.194x21.9mm, vox.dim.=0.194x0.194x1.00mm, #sl/gap=20/0.1mm, TR=’shortest’ (878.8 ±8.29ms), TE=’shortest’ (27.5 ±0.31ms),water/fat shift=27.26pix, flip=55deg, flow compensation = ‘yes’, duration=9min 11 ±5.2sec). Images were acquired with a 1D phase navigator prior to image acquisition to correct for phase errors arising from respiration during acquisition, as has been previously shown to be effective in very high resolution SWI imaging⁹.Each acquisition was processed to calculate susceptibility weighted images¹⁰, were corregistered relative to each other, and were subsequently averaged.Laminar thickness was estimated by fitting a 3rd order polynomial to all perpendicular profiles across the cortex within the FFA, and measuring the distance from edges of the cortex to the contained inflection points.

Results

Mean total cortical thickness measures of FFA2 was within the range reported in extrastriate cortex11: average FFA2 grey matter thickness of 3.52 mm, with a range between 2.5 and 4.5 mm. Laminar thickness: superficial layer thickness (1.08 mm, range: 0.71 – 1.26 mm; 31%), middle layer thickness (1.09 mm, range: 0.40 – 1.36 mm; 31%), deep layer thickness (1.35 mm, range: 0.58 – 2.42 mm; 38%).

Total FFA cortical thickness is negatively correlated with face recognition (r=-0.83 p<.001), and positively correlated with the recognition of vehicles (r=0.70, p<0.001).Face recognition was strongly correlated with the thickness of the deep cortical layers (r=-0.76, p=.002, Fig. 1B). There were no significant correlations between face recognition and thickness of the superficial or middle cortical layers (Fig. 1B). In contrast, correlations between vehicle recognition and thickness of all three laminar subdivisions were consistently positive in sign (rsranged from .44 to .48, not significant).

Conclusions

Acknowledgements

References

1. R. W. McGugin, A. E. Van Gulick, I. Gauthier, Journal of Cognitive Neuroscience 28, 282-294 (2016).

2. T. Bi, J. Chen, T. Zhou, Y. He, F. Fang, Current Biology 24, 222-227 (2014).

3. B. J. Anderson, P. B. Eckburg, K. I. Relucio, Learning & Memory 9, 1-9 (2002).

4. B. Draganski et al., Nature 427, (2004).

5. J. A. Markham, W. T. Greenough, Neuron glia biology 1, 351-363 (2004).

6. B Duchaine, K Nakayama, Neuropsychologia. 2006;44(4):576-85.

7. RW McGugin, JJ Richler, G Herzmann, M Speegle. Vision Res. 2012 Sep 15;69:10-22

8. Gauthier I1, Skudlarski P, Gore JC, Anderson AW. Nat Neurosci. 2000 Feb;3(2):191-7.

9. MJ Versluis, JM Peeters, S van Rooden, J van der Grond, MA van Buchem, AG Webb, MJ van Osch. Neuroimage. 2010 Jul 1;51(3):1082-8.

10. Haacke EM1, Xu Y, Cheng YC, Reichenbach JR. Magn Reson Med. 2004 Sep;52(3):612-8.

11. A. Parent, M. B. Carpenter, (Williams and Wilkins, Baltimore, MD, 1995).