Klaus Gramann Research Interests - Spatial Navigation


My research interest in spatial cognition focusses on the behavioral and neural foundations of spatial navigation in real and virtual space. In a series of experiments I used virtual reality (VR) to  describe the preferred use of distinct reference frames during spatial orienting (Gramann et al., 2005, JEP:HPP). In the tunnel task subjects have to adjust a homing vector indicating their end position relative to the origin of the path after traversing virtual tunnels with turns to the left or right. We demonstrated that sparse visual flow information is sufficient for accurate path integration. Moreover, subjects were found to prefer a distinct egocentric or allocentric reference frame to solve the task. “Turners” reacted as if they had taken on the new orientation during turns of the path by mentally rotating their sagittal axis (egocentric frame). “Nonturners,” by contrast, tracked the new orientation without adopting it (allocentric frame). When instructed to use their nonpreferred reference frame, both groups displayed no decline in response accuracy relative to their preferred frame; even when presented with reaction formats based on either ego-or allocentric coordinates, with format unpredictable on a trial, both groups responded highly accurately. The differences in spatial processing of identical visual flow information are not due to differences in information uptake, as indicated by comparable eye movement patterns of both strategy groups during the tunnel task (Gramann et al., 2009, International Journal of Neuroscience).  These findings support the assumption of coexisting spatial representations during navigation.

To further explore possible differences in brain dynamics accompanying the use of an egocentric or an allocentric reference frame we reconstructed the sources of brain activity during the tunnel task (Gramann et al., 2006, BrainRes). The current density reconstruction revealed the use of one or the other reference frame to be associated with distinct cortical activation patterns during critical stages of the task. For both strategy groups, an occipito-temporal network was dominantly active during the initial, straight tunnel segment (see A in Figure below). With turns in the tunnel (row B in Figure below), however, the activation patterns started to diverge, reflecting translational and/or rotational changes in the underlying coordinate systems. Computation of an egocentric reference frame was associated with prevailing activity within a posterior parietal-premotor network, with additional activity in frontal areas. In contrast, computation of an allocentric reference frame was associated with dominant activity within an occipito-temporal network, confirming right-temporal structures to play a crucial role for an allocentric representation of space.

TunnelSource
Figure 1: Source activity after clustering of relevant sources for turners (left column) and non-turners (right column) for the onset of the tunnel movement (A), the apex of turns (B), and straight segments after turns (C).

The figures display all reconstructed clusters exhibiting 75% of the maximum source activity, for 60% of the participants in a strategy group. Further investigations with patients with parietal lesions revealed the important role of the parietal cortex in encoding and recoding egocentric spatial information for further allocentric processing (Seubert, J., Humphreys, G., & Gramann, K., 2009, Neurocase).

    Currently, we are working on new methods to analyse the continuous EEG-data by means of ICA and spectral analyses (Gramann et al., 2009, Journal of Cognitive Neuroscience), replicating differences in human brain dynamics accompanying the use of distinct frames of references. Overall 35 spatially distinct clusters were derived with only 5 clusters revealing significant differences in brain dynamics between Nonturners and Turners. These clusters were located in or near the primary visual cortex, the temporo-occipital cortex, bilateral inferior parietal cortex, and the retrosplenial cortex. Using ICA on EEG data provides sufficient temporal information to investigate the time course of cognitive processing on a sub-second scale. In addition, ICA decomposes IC processes that can be localized with sufficient spatial resolution using equivalent dipole modelling.

load figure Gramann5.png

Figure 2: Equivalent model dipole locations for independent component clusters exhibiting and not-exhibiting ERSP differences between Turner and Nonturner subject groups. Blue balls show equivalent model dipole locations of independent component processes in 25 clusters without significant ERSP group differences; (green) cluster in or near right cuneus (Fig. 4B); (orange) cluster in or near bilateral inferior occipital gyrus (Fig. 4C), (red) cluster in or near left and right inferior parietal cortex (Fig. 4DE), (dark red) cluster in or near medial inferior retrosplenial cortex (Fig. 4F).

Besides differente independent frequency modes in extrastriate areas, the data revealed the retrosplenial cortex to differentiate between the use of an allocentric and an egocentric reference frame. The results support the notion that retrosplenial cortex serves as transition zone of egocentric information to an allocentric frame of reference. More importantly, we describe for the first time the sub-second dynamics of this recoding process. The following Figure displays independent component processes clustered with respect to the location of the reconstructed equivalent dipoles, their event-reltaed spectral perturbation, intertrial coherence, spectrum and scalp maps.

TunnelERSP

Figure 3: Mean event-related spectral perturbations (ERSPs) for selected independent component (IC) clusters during tunnel passages. Panel (A, left) shows baseline mean log spectra during control trials removed from the ERSPs of six selected IC clusters (B-G). Panels (A, middle) show snapshots of a representative tunnel trial at five evenly-spaced time points (spaced at intervals of 3450 ms) and (A, right) at the appearance of the response prompt. Panels (B-G, left) show locations of model equivalent dipoles for selected IC clusters, projected into a standard brain space, with each red sphere representing one cluster IC (or one of two bilaterally position-symmetric dipoles for cluster C). Panels (B-G, middle) show mean ERSP images for each of the IC clusters, revealing task-dependent changes in spectral power during navigation at log-spaced frequencies from 3 Hz to 45 Hz. Green indicates no significant difference in mean log power from baseline (visual stimulation during straight segments of the control trials). Other colors show significant deviations in log power (dB) from baseline (see color bars for scales). Vertical dashed orange lines indicate onset and offset of the period in which participants perceived the approaching and then (from 6.9 s) currently occurring tunnel turn. Vertical dashed red lines indicate the period during which subjects saw the tunnel exit approaching. (B) IC cluster 23 (22 ICs from 12 Turners and 9 Nonturners), with the centroid located in or near right cuneus (BA 18; x = -1, y = -79, z = 7; ); (C) IC cluster 21 (24 ICs from 11 Turners, 8 Nonturners), in or near bilateral inferior occipital gyrus at the border to the temporal lobe (BA19/37; x = 37, y = -70, z = -1); (D) IC cluster 17 (26 ICs, 9 Turners, 10 Nonturners) in or near precuneus (BA 7; x = 0, y = -45, z = 43); (E) IC cluster 12 (45 ICs, 12 Turners, 11 Nonturners) in or near the right precentral gyrus (BA 4; x = 36, y = -12, z = 49); (F) IC cluster 8 (26 ICs, 7 Turners, 8 Nonturners) in or near the left precentral gyrus (BA 4; x = -40, y = -13, z = 44); (G) IC cluster 1 (24 ICs, 13 Turners, 9 Nonturners) in or near the right medial frontal gyrus (BA 9; x = 2, y = 41, z = 26). See Supplemental Figures 1-6 for a description of all IC clusters.

load image Figure4Gramann.png

Figure: Component clusters revealing significant differences for the use of an egocentric or an allocentric reference frame during spatial navigation. Top row (A) displays the time course of a representative tunnel trial with five segments for Nonturners, using an allocentric reference frame, for Turners, using an egocentric reference frame, and for the difference ERSPs (computed by subtracting ERSP activity of Turners from ERSP activity o Nonturners). Rows B to F display from the left to the most right column frequency-rescaled i) baseline power spectrum of control trials, ii) ERSP activity for Nonturners, iii) ERSP activity for Turners, iv) difference ERSPs computed by subtracting ERSP activity of Turners from ERSP activity of Nonturners, and v) cluster IC equivalent dipoles projected on horizontal, sagittal, and coronal view of the standard brain for (B) a component cluster with the cluster centroid located in or near the right inferior occipital gyrus (x = -37, y = -70, z = -1), (C) a component cluster in or near bilateral inferior occipital gyrus at the border to the temporal lobe (BA19/37; x = 37, y = -67, z = -1), (D) a component cluster located in or near left inferior parietal lobule (BA 40; x = -30, y = -40, z = 33), (E) a component cluster located in or near the right precuneus (x = 26, y = -47, z = 32), and (F) a component cluster located in or near the retrosplenial cortex at the posterior pole of the cingulate cortex (x = 1, y = -56, z = 10). Color coding and dashed vertical lines as in Figure 2.

    Our results demonstrate that advanced EEG-based functional brain imaging using ICA-derived spatial filtering can be used to study network dynamics of spatial orienting and navigation. Using data-driven ICA spatial filtering on high-density EEG data and equivalent-dipole modeling of IC source locations, we are able to describe the task- and strategy-dependent modulation of several frequency bands underlying computation and maintenance of distinct reference frames on a sub-second time scale. Also, our results show that future behavioral and brain imaging studies of human navigation should take into account individual subject differences in navigation approach and strategy. Our results support our conclusion that preferred use of an egocentric or an allocentric reference frame during spatial navigation is accompanied by differences in EEG brain dynamics in cortical areas involved in integrating visual flow information with changes in cognitive heading, and in areas associated with the transfer of egocentrically experienced spatial information into an allocentric reference frame.

Link to all reconstructed IC clusters for the above described experiment (Supplementary Material).
Link to the Tunnel paradigm.
back
Department Psychology, Ludwig-Maximilians-University Munich
Swartz Center for Computational Neuroscience, UCSD



Link to Tunnel download and current work in progress using the Tunnel paradigm.
back
Department Psychology, Ludwig-Maximilians-University Munich
Swartz Center for Computational Neuroscience, UCSD