Understanding space by moving through it: neural networks of motion- and space processing in humans

Humans explore the world by moving in it, whether moving their
whole body as during walking or driving a car, or moving their arm
to explore the immediate environment. During movement, self-motion
cues arise from the sensorimotor system comprising vestibular,
proprioceptive, visual and motor cues, which provide information
about direction and speed of the movement. Such cues allow the body
to keep track of its location while it moves through space.
Sensorimotor signals providing self-motion information can
therefore serve as a source for spatial processing in the brain.
This thesis is an inquiry into human brain systems of movement and
motion processing in a number of different sensory and motor
modalities using functional magnetic resonance imaging (fMRI). By
characterizing connections between these systems and the spatial
representation system in the brain, this thesis investigated how
humans understand space by moving through it. In the first study of
this thesis, the recollection networks of whole-body movement were
explored. Brain activation was measured during the retrieval of
active and passive self-motion and retrieval of observing another
person performing these tasks. Primary sensorimotor areas dominated
the recollection network of active movement, while higher
association areas in parietal and mid-occipital cortex were
recruited during the recollection of passive transport. Common to
both self-motion conditions were bilateral activations in the
posterior medial temporal lobe (MTL). No MTL activations were
observed during recollection of movement observation. Considering
that on a behavioral level, both active and passive self-motion
provide sufficient information for spatial estimations, the common
activation in MTL might represent the common physiological
substrate for such estimations. The second study investigated
processing in the 'parahippocampal place area' (PPA), a region in
the posterior MTL, during haptic exploration of spatial layout. The
PPA in known to respond strongly to visuo-spatial layout. The study
explored if this region is processing visuo-spatial layout
specifically or spatial layout in general, independent from the
encoding sensory modality. In both a cohort of sighted and blind
participants, activation patterns in PPA were measured while
participants haptically explored the spatial layout of model scenes
or the shape of information-matched objects. Both in sighted and
blind individuals, PPA activity was greater during layout
exploration than during object-shape exploration. While PPA
activity in the sighted could also be caused by a transformation of
haptic information into a mental visual image of the layout, two
points speak against this: Firstly, no increase in connectivity
between the visual cortex and the PPA were observed, which would be
expected if visual imagery took place. Secondly, blind
participates, who cannot resort to visual imagery, showed the same
pattern of PPA activity. Together, these results suggest that the
PPA processes spatial layout information independent from the
encoding modality. The third and last study addressed error
accumulation in motion processing on different levels of the visual
system. Using novel analysis methods of fMRI data, possible links
between physiological properties in hMT+ and V1 and
inter-individual differences in perceptual performance were
explored. A correlation between noise characteristics and
performance score was found in hMT+ but not V1. Better performance
correlated with greater signal variability in hMT+. Though
neurophysiological variability is traditionally seen as detrimental
for behavioral accuracy, the results of this thesis contribute to
the increasing evidence which suggests the opposite: that more
efficient processing under certain circumstances can be related to
more noise in neurophysiological signals. In summary, the results
of this doctoral thesis contribute to our current understanding of
motion and movement processing in the brain and its interface with
spatial processing networks. The posterior MTL appears to be a key
region for both self-motion and spatial processing. The results
further indicate that physiological characteristics on the level of
category-specific processing but not primary encoding reflect
behavioral judgments on motion. This thesis also makes
methodological contributions to the field of neuroimaging: it was
found that the analysis of signal variability is a good gauge for
analysing inter-individual physiological differences, while
superior head-movement correction techniques have to be developed
before pattern classification can be used to this end.