Processing of biological movements by the human brain

Introduction

For human beings, the movements of living creatures are an important and at times vital source of information. Biological movements can be used to identify moving things as alive, can tell us about what goals they are pursuing or what they feel and thus help us decide how to interact with them. However, between simply detecting “aliveness” or animacy in a moving object and processing and categorizing subtle facial movements, there is a whole range of processing at work and little is known about the brain mechanisms involved, and even less is known about their dysfunctions.

Goals

Our goal is to understand how the human visual system processes biological movements, in their simplest to most complex manifestations. Specific goals include finding which brain regions are involved in general and which are specific to particular kinds of biological motion, how the different kinds of information conveyed by face motion are processed by the face processing network, and how subtle but behaviorally relevant differences in facial motion are differentiated and represented.

Methods

We take great care to use highly controlled, parametric stimuli in all our experiments. For example, we have created movement algorithms that make a single dot appear animate or not with minimal context information, while keeping low-level motion characteristics almost identical [Figure 1A]. Regarding facial motion, we parametrically varied frame rate and frame order to separate static information from meaningful, fluid motion [Figure 1B]. These stimuli are used in classical psychophysical and functional brain imaging experiments. Representations of biological movements with a high variability are studied using multivariate analysis methods applied to both psychophysical and brain imaging data. While we mostly study the healthy population, we start looking at deficits in detection of animacy and social interactions from biological movements in Autism. Students and researchers at the MPI and in several other institutions collaborate in these projects.

Figure 1 A) Animacy judgments about a single moving dot as a function of the motion parameter controlling the dots movement equation. Black lines show mean ratings over all subjects and fitted cumulative gaussian, grey lines show ratings of each subject (N=20). B) Perceived fluidity of the facial motion in movies presented at different frame rates, with frames either in correct or in scrambled order (mean over 10 subjects). The latter stimuli are controls with the same amount of static information but no fluid motion.

Initial results and conclusions

Our results indicate that the superior temporal sulcus (STS) plays a central role in processing most kinds of biological motion as expected from previous work [Figure 2A], with the notable exception of animate-looking single moving objects. We found that facial motion boosts the response of face-selective regions involved in processing face identity [1,2] [Figure 2B], prompting a follow-up study asking whether idiosyncratic facial motion carrying identity information is directly analysed in these regions. Another new finding is that frontal regions show a stronger categorical response than STS to animate-looking moving objects [3] [Figure 2C], disproving the hypothesis that STS is a universal life-detector. Lastly, results from our Autism study reveal dysfunctions in identification of interacting moving objects, but not in processing isolated, animate-looking objects [4]. We are currently working on confirming and following up on these results and are working on updates of current theories to fit our findings.

Figure 2 A) Brain regions with BOLD response increasing with fluidity of facial motion are shown in red. Strongest and widest response is observed in STS. B) BOLD response in the fusiform face area of the right hemisphere to facial motion stimuli with different frame rates and frame orders. Results show a stronger response to stimuli with multiple frames (5, 12.5 and 25 Hz) than to a single frame (1 Hz), with a stronger response to ordered than to scrambled frames at higher frame rates. C) Cluster of voxels showing a categorical response to animacy stimuli.

References

1.  Schultz J and Pilz KS (2009) Natural facial motion enhances cortical responses to faces. Experimental Brain Research 194(3) 465-475.

2.  Schultz J, Brockhaus M, Bülthoff HH and Pilz KS (2011) What the human brain likes about facial motion. Submitted.

3.  Schultz J and Bülthoff, HH (2011) Brain regions involved in detection of animacy from a single moving object. In preparation.

4.  David N, Schultz J, Milne E, Schunke O, Schöttle D, Münchau A, Vogeley K, Siegel M, Engel AK (2011) Selective alteration of social-interactive motion signal detection in autism spectrum disorders. In preparation.

Seit Oktober 2010 - jetzt: co-Projektleitr der Gruppe Erkennen und Kategorisieren in Prof. Bülthoff's Abteilung.

Oktober 2004 - jetzt: post-doctoral Research Scientist in Prof. Bülthoff's Abteilung am the Max Planck Institut für Biologische Kybernetik. Finanzierung: Max Planck Gesellschaft.

2004: Ph.D. in Kognitiven Neurowissenschaften am Wellcome Department of Imaging Neuroscience der University College London, Grossbritannien. Thema: Wahrnemung komplexer Bewegungen bei Menschen, studiert mit funktionaler Kernspintomographie und Psychophysik. Betreuer: Chris D. Frith und Daniel M. Wolpert.

2000: Diplom in Medizin an der Universität Genf, Schweiz. Zusätzliche Kurse und Ausbildung in Experimentalpsychologie, Neuropsychologie und Molekularbiologie.