INTRODUCTION

Our understanding of human somatosensory cortex is limited compared to other, more well-studied sensory cortices (e.g., visual and auditory regions). One reason for this is that primary somatosensory cortex (S1) is a relatively small region, and hence more difficult to non-invasively access in awake and behaving humans. Although there have been PET studies of S1 processing¹ dating back to the 1980’s as well as conventional resolution fMRI studies² in the subsequent couple decades, it wasn’t until the advent of high-resolution (sub-millimeter) fMRI that it became possible to map the finer details of human S1 – such as the somatotopic representations of individual body parts (e.g., fingertips)³. Another factor limiting somatosensory research has been the difficulty associated with constructing well-controlled and appropriate stimuli. Natural objects vary across many dimensions making it difficult to compare responses to different materials, and simple man-made stimuli such as tactile gratings often lack ecological validity. Recent advances in 3D-printing offers a way forward by permitting the generation and modification of a wide variety of textures and objects - all using MR-safe materials.

Natural forms, often characterized by irregularity and roughness, have a unique complexity that exhibit self-similarity across different spatial scales or levels of magnification⁴. This fractal-like scaling characteristic is ubiquitous in many physical and biological domains, and recent research has indicated that the cortex – at least in response to visual stimuli – appear to be tuned for the particular fractal structure found in nature⁵. Despite both the high ecological validity of fractal structures and the prominent role that the scale-specific processing plays in various sensory modalities, the fractal-scaling framework is largely absent in both research and theory of perceptual processing and experience. Here we aimed to investigate the possible utility of 3D-printed fractal textures for probing processes in human somatosensory cortex using 7T fMRI.

METHODS

Data were acquired on a MAGNETOM 7T whole-body research scanner (Siemens Healthcare) with a 32-channel head coil. Whole-brain anatomical images were collected using an MP2RAGE sequence with an isotropic voxel size of 0.5mm. All functional data were collected using a 3D-EPI sequence⁶ with an isotropic voxel size of 0.8mm and an effective volume TR of 2s. Models of the 3D tactile stimuli⁴ were generated following the three steps illustrated in Figure 1. First, a series of 2D, 1/f noise images with varying amplitude spectra slopes were created (top row). Next, to create 3D solids (middle row), the tonal values of each pixel in the image was converted to a height value in three-dimensional space. These models were then 3D printed onto one face of a 10 x 10 x 1cm hard, MR-safe, synthetic block (bottom row). Note that the scale invariance of natural scenes can be captured by a geometric scaling parameter known as the fractal dimension (D). The D values of the textures used in our MRI-experiment were: 1.25, 1.50, 1.75, 2.00, 2.25, and 2.50.

Each subject (n=3) underwent two scan sessions: (1) somatotopic digit mapping and (2) the fractal touch experiment. For digit mapping: the 4 fingertips (index, middle, ring, and little) on the right hand were stimulated sequentially using a phase-encoded design⁷. To this end, a single fingertip was stimulated using a piezoelectric stimulator (Figure 2A) for 8s before moving to the next. Stimulation began at the first finger (index) and ended with the fourth finger (little). Stimulation then returned to the first finger to begin another cycle of stimulation. The frequency of vibrotactile stimulation changed every 2s, and three different frequencies were used (5, 20, and 100Hz). For the fractal touch experiment, the 6 different textured stimuli were affixed around the edge of a disk (Figure 2B). During each run, the participant was prompted by visual cues to stroke one of the textured stimuli using all four fingers in an ON/OFF manner. The participant then rotated the disk clockwise or counter-clockwise and repeated until all 6 textured stimuli had been used. Four runs were collected, two clockwise and two counter-clockwise – presented in alternating order. The order of the stimuli was randomized across participants.

RESULTS

The 3D-printed fractal textures were found to be suitable and effective for use in the MR-environment. Simple stroking of the blocks elicited robust signals throughout primary somatosensory cortex for each participant and these overlapped with the individual digit representations defined using the data from the somatotopic mapping condition (Figure 3). Despite the robust activation elicited by stroking the textured stimuli, we found no evidence at the individual level that the BOLD activity in S1 was sensitive to the fractal dimension of the stimuli. Instead, we found that all stimulus textures elicited similar levels of activation (Figure 4).

DISCUSSION

This study shows that 3D-printed fractal textures can be used to elicit strong and wide-spread activity across the digit representations in primary somatosensory cortex. The strength of activation was found to be similar across all textures suggesting that S1, or at least the BOLD signal therein, is not sensitive to the fractal dimension within the range tested. Although we demonstrate its utility for generating fractal textures, 3D-printing is flexible and well-suited for producing a range of other interesting textures and objects for use in the MR-environment.

Acknowledgements

We thank Aiman Al-Najjar, Nicole Atcheson, Saskia Bollmann, and Markus Barth for help with data collection, and the authors acknowledge the facilities of the National Imaging Facility at the Centre for Advanced Imaging, University of Queensland. This work was supported by the Australian Research Council (DE180100433) as well as the National Health and Medical Research Council (APP 1088419).