Direct evidence for encoding of motion streaks in human visual cortex

Abstract

Temporal integration in the visual system causes fast-moving objects to generate static, oriented traces ('motion streaks'), which could be used to help judge direction of motion. While human psychophysics and single-unit studies in non-human primates are consistent with this hypothesis, direct neural evidence from the human cortex is still lacking. First, we provide psychophysical evidence that faster and slower motions are processed by distinct neural mechanisms: faster motion raised human perceptual thresholds for static orientations parallel to the direction of motion, whereas slower motion raised thresholds for orthogonal orientations. We then used functional magnetic resonance imaging to measure brain activity while human observers viewed either fast ('streaky') or slow random dot stimuli moving in different directions, or corresponding static-oriented stimuli. We found that local spatial patterns of brain activity in early retinotopic visual cortex reliably distinguished between static orientations. Critically, a multivariate pattern classifier trained on brain activity evoked by these static stimuli could then successfully distinguish the direction of fast ('streaky') but not slow motion. Thus, signals encoding static-oriented streak information are present in human early visual cortex when viewing fast motion. These experiments show that motion streaks are present in the human visual system for faster motion.

Figure 1.

Motion streaks in art and vision. (a) Motion streaks are often used in photography and art to give a strong impression of fast motion within a scene. Photograph by Tod Klassy, sourced from www.flickr.com and reproduced with permission. (b) Geisler's [1] model of how a motion streak might be combined with a motion signal in early cortex to provide a code for motion direction. Specifically, a direction-selective V1 cell (giving the sign of motion direction) might combine its output with that of a cell selective for static orientation, which would respond to the temporally integrated motion streak, giving fine angular resolution and solving the aperture problem [1].

Figure 2.

Schematic of the procedure for the psychophysical experiment. Participants adapted to vertical motion, either faster (13 m s⁻¹) or slower (1.6 m s⁻¹) than the streak threshold for dots of this size. They were then asked to detect a low-contrast grating, either parallel (a) or orthogonal (b) to the direction of motion, which appeared either to the left or to the right of fixation, in the same retinal location as the adapting dots. Contrast of the test grating was controlled by a QUEST adaptive staircase; see §2 for full details.

Figure 3.

Mean results from the psychophysical adaptation experiment for eight participants. Error bars denote ±1 s.e.m. There was a significant interaction between orientation and speed, F_1,7 = 39.29, p < 0.001; see §§3 and 4 for details. Black bars, parallel; grey bars, orthogonal.

Figure 4.

Procedure for the fMRI experiment and univariate results. (a) Schematic of the block design within each scanning session. (b) Statistical parametric maps from a single representative participant overlaid on a three-dimensional reconstruction of a T1 template brain in the stereotactic space of Talairach & Tournoux [29]. Red colours indicate those cortical loci that showed greater BOLD responses to faster compared with slower motion. A threshold of p < 0.001 (uncorrected) is used for display purposes. Green regions showed greater responses to slow motion than oriented stimuli (p < 0.001). (c) Mean per cent BOLD signal change (relative to global mean) in each region of interest, averaged over eight participants. Error bars denote ±1 s.e.m.

Figure 5.

Mean decoding accuracy across eight participants for five regions of interest. (a) Results for decoding the orientation of static stimuli (45° versus 135°) for stimulus-responsive regions in early retinotopic visual cortex plus V5/hMT+ and a control region in prefrontal cortex. (b,c) Results for decoding of faster and slower motion. (d,e) Training the classifier on discriminating orientation but testing it by discriminating the direction of faster motion (d) or vice versa (e). (f) Training the classifier to discriminate the orientation, but testing it on slower motion. (g,h) Generalizations for faster-to-slower and slower-to-faster motion. (i) Generalization from slower motion to orientation. The dashed line indicates chance performance, and the shaded region indicates its 95% CI (see §2 for details). Error bars denote ±1 s.e.m. Asterisks indicate regions where decoding accuracy was significantly (p < 0.01, two-tailed t-test) different from chance performance.