Project Leader

Dr. Andreas Bartels
Phone: +49 7071 601-656
Fax: +49 7071 601-652
andreas.bartels[at]tuebingen.mpg.de

 

Visual Perception - Psychophysics, Physiology and fMRI Studies

Distinct regions in the human brain respond preferentially to global (self-motion) or local (object-motion) cues. (a) Brain activity related to global and local motion when humans view a movie stimulus; (b): fMRI signal response amplitudes of the regions in (a); (c): sample frames of the movie stimulus that contain primarily local or global motion, respectively. Superimposed are flow vectors (in red) obtained from the computational analysis of the movie [3].
Distinct regions in the human brain respond preferentially to global (self-motion) or local (object-motion) cues. (a) Brain activity related to global and local motion when humans view a movie stimulus; (b): fMRI signal response amplitudes of the regions in (a); (c): sample frames of the movie stimulus that contain primarily local or global motion, respectively. Superimposed are flow vectors (in red) obtained from the computational analysis of the movie [3].
Humans are above all visual beings. Vision allows us to quickly obtain information about a plethora of aspects of our world: the spatial 3D layout of our surroundings, the distance to objects within grasping distance, the reflectance properties of surfaces, our own velocity and that of objects around us, the identities and even emotions of other people – all this within the blink of an eye. Accordingly, no sensory loss is more debilitating than blindness. It is thus no coincidence that a prime interest in neuroscience revolves about understanding the function of the neural substrates serving the visual sense, which occupy an astonishing 30 to 40 percent of the cerebral cortex’ total surface area. One hope is that insights gained from the study of the visual system might be transferred to help understand other neural systems in the brain. For these reasons a major effort of our department is aimed at elucidating a few crucial aspects of visual function – using psychophysics, electrophysiology, EEG as well as fMRI.

In real life, visual cues hardly ever stand still for two reasons: our eyes, head and body lead to nearly incessant motion cues and, independently, there is “real” motion of objects around us. We are investigating visual motion and self-motion processing using human fMRI to identify and characterize the properties, connectivity and causal involvement of various higher-level motion processing regions in the human brain that separate these distinct motion signals, allowing us to perceive a stable world, and that exploit the cues of dynamic perspective changes to inform us about the 3D layout of our environment (see also Figure 1).

The perceptual richness of dynamic natural visual stimuli is not only suited to address questions of function, but also those pertaining to the nature of the signals that neuroscientists measure in the brain. by predicting spiking activity from non-invasive EEG in behaving macaques, we exploit the complex response to natural stimuli in order to relate neurophysiological signals obtained invasively with those acquired non-invasively. From this, we can gain a better understanding of how well such non-invasive signals reflect local neuronal activity, which is of great relevance to both clinical and basic research in humans.

Colour information complements that of motion cues, as the two are entirely independent of each other. Colour and motion have also been shown to be processed in largely independent pathways in the cortex. This gives rise to the “binding problem”: how do we perceive distinct visual cues in apparently perfect co-registration when they are processed independently? To investigate the coding of colour, motion, orientation and of their conjunctions in the visual brain is one of our attempts to address the problem of feature integration within the visual system in the human brain.

In addition to studying neural responses that are evoked by changes in physical stimuli, it is a fundamental aim of neuroscience to relate subjective perceptual experiences to objective measures of neural activity. Bi-stable percepts in general and binocular rivalry in particular provide an ideal starting point for doing so, because here the subjective percepts change over time even though the physical stimulation remains constant. In binocular rivalry two dissimilar images are presented to the two eyes, which leads to percepts that alternate between the two. The interest in this phenomenon from a neuroscientific perspective is straightforward: if we understand which neural processes are responsible for selecting the perceived (dominant) stimulus representation rather than the non-perceived (suppressed) one, we may be one step closer to understanding the mechanisms underlying conscious visual perception. There are additional reasons that make the study of binocular rivalry interesting: the stochastic nature of the perceptual transitions in rivalry resembles that of many other bi-stable perceptual phenomena, such as the vase-face illusion, or those arising in other modalities such as the tactile or auditory. Perceptual transitions and the stable phases in between may thus be indicators of dynamic properties underlying sensory processing in general. An understanding of the neural processes involved in rivalry may thus also shed light on fundamental properties of neural processing [1, 2].
Extracellular recordings in alert behaving monkeys show that modulations of mean population activity following a perceptual switch are significantly lower in the macaque primary visual cortex (a) compared to ventrolateral prefrontal cortex (b). Interestingly, scalp EEG experiments in humans demonstrate that the anterior prefrontal cortex is the source of an increase in the 12 – 30Hz frequency range following a perceptual switch (c). (Arrows in plots a, b and c indicate the occurrence of a perceptual switch. Note that in the periods after a perceptual switch sensory stimulation is kept constant, and any differences reflect perceptual effects.)
Extracellular recordings in alert behaving monkeys show that modulations of mean population activity following a perceptual switch are significantly lower in the macaque primary visual cortex (a) compared to ventrolateral prefrontal cortex (b). Interestingly, scalp EEG experiments in humans demonstrate that the anterior prefrontal cortex is the source of an increase in the 12 – 30Hz frequency range following a perceptual switch (c). (Arrows in plots a, b and c indicate the occurrence of a perceptual switch. Note that in the periods after a perceptual switch sensory stimulation is kept constant, and any differences reflect perceptual effects.)
The importance of the above questions leads us to study the problem at distinct levels of the visual processing hierarchy using a variety of methodological approaches (see Figure 2). We are investigating the mechanisms of perceptual suppression in primary visual cortex to reveal the small, but potentially highly important perceptual modulations at the earliest stages of visual processing. It is currently unclear whether these originate locally or as a result of feedback from higher or highest levels of processing, where much stronger perceptual modulations are observed, as an investigation of the neurophysiological mechanisms of binocular rivalry in the macaque prefrontal cortex shows. A partial answer to this is provided by the psychophysical evidence in binocular rivalry experiments with a time-dependence of eye and stimulus contributions. This investigation demonstrates that the strength of monocular, i.e. eye-based contributions to the percept co-varies precisely with the strength of image representations, suggesting that the two may be coupled via feedback.

Comparing the interocular switch and classical binocular rivalry in the human brain using EEG provides evidence of differences in stable perception while viewing of non-ambiguous or ambiguous stimuli and, interestingly, identifies differences in oscillatory power in the pre-frontal cortex in these conditions.

Finally, we try to gain further insight into the mechanisms of bistable perception by using neuronal data to constrain computational models. In such a neuronal model we find evidence that a perceptual switch is induced by an optimal combination of noise and spiking frequency adaptation. These studies provide first clues to the mechanisms underlying perceptual switches and the maintenance of stable percepts.

By pinning down the neural correlates of consciousness, we investigate the extent to which physiological changes previously attributed to perceptual changes are or are not affected by pharmacological manipulation that removes conscious states altogether. It is a fundamental challenge not only to establish psychophysical and physiological correlates of perception, but also to distinguish between neural processes that cause it and those that are a consequence of it. Our research may bring us one step closer to understanding how the brain represents the world around us.

References

1. Logothetis, N. K., D. A. Leopold, D. L. Sheinberg: What is Rivaling during Binocular Rivalry? Nature 380, 621–624 (1996).
2. Blake, R., N. K. Logothetis: Visual Competition. Nat Rev Neurosci 3, 13–21 (2002).
3. Bartels, A., S. Zeki, N. K. Logothetis: Natural Vision Reveals Regional Specialization to Local Motion and to Contrast-invariant, Global Flow in the Human Brain. Cereb. Cortex 18, 705–717 (2008).
Last updated: Wednesday, 26.10.2011