Edited by: Hong-Jin Sun, McMaster University, Canada
Reviewed by: Brett R. Fajen, Rensselaer Polytechnic Institute, USA; Albert Van Den Berg, University Medical Centre St. Radboud Nijmegen, Netherlands
*Correspondence: Jac Billington, Institute of Psychological Sciences, Faculty of Medicine and Health, The University of Leeds, Leeds LS2 9JT, UK. e-mail:
This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in other forums, provided the original authors and source are credited and subject to any copyright notices concerning any third-party graphics etc.
Visual control of locomotion typically involves both detection of current egomotion as well as anticipation of impending changes in trajectory. To determine if there are distinct neural systems involved in these aspects of steering control we used a slalom paradigm, which required participants to steer around objects in a computer simulated environment using a joystick. In some trials the whole slalom layout was visible (steering “preview” trials) so planning of the trajectory around future waypoints was possible, whereas in other trials the slalom course was only revealed one object at a time (steering “near” trials) so that future planning was restricted. In order to control for any differences in the motor requirements and visual properties between “preview” and “near” trials, we also interleaved control trials which replayed a participants' previous steering trials, with the task being to mimic the observed steering. Behavioral and fMRI results confirmed previous findings of superior parietal lobe (SPL) recruitment during steering trials, with a more extensive parietal and sensorimotor network during steering “preview” compared to steering “near” trials. Correlational analysis of fMRI data with respect to individual behavioral performance revealed that there was increased activation in the SPL in participants who exhibited smoother steering performance. These findings indicate that there is a role for the SPL in encoding path defining targets or obstacles during forward locomotion, which also provides a potential neural underpinning to explain improved steering performance on an individual basis.
A crucial element of survival for most animals is the ability to move through their environment successfully; steering toward objects of interest (e.g., food) and avoiding collisions with dangerous objects (e.g., a predator or concrete barrier). Such locomotor tasks require the integration of several informational variables available within the visual scene (Wilkie and Wann,
Research in both non-human primates and humans has revealed a network of cortical regions which show preferences toward global optical flow components which are indicative of self motion and provide valuable information regarding current heading. MST, a sub-region of the human motion complex (MT+) located in the superior temporal cortex has been proven to show robust activation to visual cues which are compatible with self motion (global expansion and rotation patterns) in primates (Duffy and Wurtz,
Several studies have examined human cortical involvement whilst making judgments of heading. The earliest of these, by Peuskens et al. (
The evidence discussed so far highlights functional cortical regions associated with optic flow stimuli consistent with egomotion. Very few studies, however, have attempted to simulate visual-motor scenarios more akin to natural locomotor steering, where objects serve to delineate the desired future path. Field et al. (
The studies of Billington et al. (
Locomotion through the environment is not always guided by a continuously demarked path. Alternatively, locomotion can involve continuously updating and predicting the future location of obstacles in relation to the self, and each other, in order to pursue a self-initiated pathway. A study by Wolbers et al. (
This study aims to build upon our previous research regarding neural contributions to effective heading detection and effective steering along predetermined pathways (Field et al.,
Fourteen neurotypical participants (10 female, 4 male) between 20 and 36 years of age (mean 28.31,
Stimuli were presented to the participant via a NordicNeuroLab VisualSytem© with integrated optical diopter correction (−5pt to +2pt). The OLED display had 30° horizontal × 23° vertical display (800 × 600 pixels), all of which was visible to the participants. This system also allowed the monitoring of eye movements during trials. Participants were asked to lie comfortably in the scanner and had the VisualSystem lowered onto their eyes. Interpupillary distance was measured in order to set the optimum goggle disparity and diopter correction was used on participants requiring corrective eyewear.
Each condition was visually matched in that the lower half of the vertical axis contained a textured ground plane which provided optic flow cues as participants moved though the scene, and the upper half of the vertical axis contained a blue sky plane (see Figure
The general scene displayed for all conditions was of the ground plane strewn with numerous yellow cones placed at random locations (0.0079 cones/m2). This presented a cluttered scene with multiple object features but participants did not have to directly attend to these cones to complete the task.
Each trial block lasted 20 s with a between block rest duration of 7.9–8.5 s (random uniform distribution) in which a blue blank screen was presented. Each condition was presented 10 times in total over two separate runs. Five seconds before the start of each block a text cue appeared centrally on this screen instructing the participants as to the task that would appear in the following block.
If the text “Steer” was shown in the pre-cue period participants were presented with one of two conditions:
If the text “Passive” was shown in the pre-cue period participants were presented with one of three conditions, which matched the visual content of the steering trials:
A series of cones were visible in this condition; however, they were randomly placed and not synced with the heading trajectory and therefore provided no indication about forthcoming changes in heading. This condition is a well matched control to SteerPv in terms of visual information and motor response; however, there are no requirements to plan steering responses using information regarding the spatial location of cones.
We used a saccadic eye movement task to localize the parietal region thought to be the human homologue of the lateral intraparietal area in monkeys. A similar saccadic eye movement task was found to be an effective PEF localizer in our previous study and was used as an exclusive mask in order to remove cortical activation resulting from low level attentional effects and eye movements (Billington et al.,
We presented participants with alternating Sacc and Fix blocks (16 s) and gave instructions to follow the dot on the screen at all times. The localizer lasted 256 s, with eight repetitions of each Sacc and Fix block. During Fix blocks the dot remained stationary in the center of the screen. During the Sacc block the dot position was randomly updated every 500 ms. The maximum horizontal eccentricity of the dot was 12.5° from the center of the screen and the maximum vertical eccentricity was 6.25° from the center of the screen.
Behavioral steering data was collected at 60 Hz using an MRI compatible joystick (MAG Concept, Redwood City, CA). This joystick was placed to the right hand side of the participant on the scanner bed so it was possible to steer comfortably with the joystick in the right hand for the duration of the scanning session (only right-handed participants were used). The maximum possible turning speed was 40.91°/s when the joystick was fully engaged to the left or right [participants used 14.18°/s (
Eye tracking data was collected via an integrated NordicNeuroLab eye tracking camera (60 Hz) using Arrington software (Arrington Research, Scottsdale, AZ). Eye calibration grids were presented before both slalom runs and this data was used to standardize the data from each participant's slalom runs. This involved converting
fMRI data were collected using a Siemens Trio 3 Tesla scanner with an eight-channel head array coil. Functional images were collected using 38 slices covering the whole brain (slice thickness 3 mm, interslice distance 0 mm, in-plane resolution 3 × 3 mm) with an echo planar imaging sequence (
Individual statistical contrasts were set up using the general linear model to fit each voxel with a combination of functions derived by convolving the standard haemodynamic response with the block design time series. Six additional regressors were added to each model in order to model potentially confounding rotational and translational minor head movements in
To steer a slalom participants must continually modify locomotor heading (°) giving rise to a change in angular velocity (°/s). Steering a smooth sinusoidal path would result in changes in aAcc (°/s2), whereas abrupt changes to the trajectory would result in increased angular jerk (°/s3). Mean values for both aAcc (aAcc) and angular jerk (aJrk) are shown in Table
HeadingNr | 4.554 | 2.182 | 227.707 | 123.555 | 0.889 ( |
0.470 |
HeadingPv | 4.504 | 2.429 | 217.993 | 99.297 | 0.888 ( |
0.500 |
SteerNr | 4.718 | 2.407 | 233.04 | 126.888 | ||
SteerPv | 4.467 | 2.316 | 233.0794 | 146.227 | ||
BL | 4.630 | 2.118 | 231.832 | 118.097 | 0.916 ( |
0.480 |
No differences in aAcc and aJrk measures were found between the heading and steering trials. This is consistent with the fact that heading trials were replays of previous steering trials and therefore would elicit similar amplitude joystick movements from participants. Heading trials were associated with ~0.9 s lag in joystick response to on screen heading (see Table
In order to ascertain whether there was any cortical activation present as a result of just seeing red and blue cones either continuously on the screen, or fading in as the trial progressed we compared both passive conditions, HeadingNr and HeadingPv, to the BL condition. Activation was present in parietal regions (bilateral precuneus; see Table
HeadingNr > BL | Precuneus | −23 | −67 | 30 | 198 | 4.0 | <0.001 (unc.) |
−4.9 | −55 | 49 | 577 | ||||
6.4 | −53 | 45 | 441 | ||||
HeadingPv > BL | Precuneus | −21 | −64 | 52 | 1179 | 4.0 | <0.001 (unc.) |
5.8 | −56 | 45 | 324 | ||||
SteerNr > HeadingNr | Cingulate gyrus | −9 | −24 | 42 | 301 | 3.861 | <0.05 (FDR) |
Superior parietal lobe | −9.6 | −65 | 51 | 1782 | |||
−9 | −72 | 47 | 274 | ||||
Lateral occipitotemporal gyrus | −36 | −67 | −13 | 95 | |||
SteerPv > HeadingPv | Cingulate sulcus | −8.9 | −23 | 43 | 1157 | 3.379 | <0.05 (FDR) |
Central sulcus | 22 | −28 | 55 | 734 | |||
−15 | −30 | 61 | 454 | ||||
Post central sulcus | 26 | −40 | 47 | 1121 | |||
Superior parietal lobe | 13 | −59 | 58 | 721 | |||
−24 | −50 | 56 | 1491 | ||||
−12 | −59 | 52 | 628 | ||||
Precuneus | 9 | −59 | 55 | 1383 | |||
−9 | −48 | 49 | 998 | ||||
Ventral intraparietal area | 25 | −66 | 33 | 1207 | |||
Middle occipital gyrus | 36 | −76 | 12 | 469 | |||
Medial occipitotemporal gyrus | 31 | −33 | −17 | 912 | |||
Lingual gyrus | 10 | −31 | −11 | 478 |
Table
An ANCOVA analysis was carried out in Brain Voyager in order to determine whether the differences in steering smoothness in individual participants contributed to differences in cortical activation. For this we explored the variation between the standardized aAcc and aJrk values and the activation (in terms of mean BOLD response) in either the SteerPv or SteerNr trials (
In order to rule out the possibility that the aforementioned activations in SPL and postcentral sulcus merely reflected either individual variations in some mechanical aspects of steering with a joystick, or low level visual aspects of screen motion we examined aAcc and aJrk correlations during passive heading trials. The same regression analysis between aAcc and aJrk values and activation during HeadingNr and HeadingPv (vs. BL) did not reveal any cortical activity in associated with aAcc or aJrk scores. For heading trials lag values were regressed with activation in HeadingNr and HeadingPv (vs. BL) to determine cortical regions associated with maintaining timely joystick movements to on screen heading. Despite obvious individual differences in these scores (as highlighted by large
Collated group eye tracking data are presented in two dimensional “heat maps” (Figure
To our knowledge this is the first study to investigate how the human brain encodes and updates target locations and uses this visual information for the purpose of steering through an obstacle rich environment. This study also reveals that key cortical regions are differentially activated according to the smoothness of the steering trajectory.
In this study we were particularly interested in the behavioral and neural responses related to advance planning of the trajectory. When participants were presented with the whole slalom course at the start of the trial (SteerPv) they were able to plan ahead and so we would expect smooth trajectories, whereas when only the nearest slalom objects were visible (SteerNr) last minute responsive changes would have to be executed without future planning. It seems clear that near information should be more useful for immediate error correction whereas distant information about future waypoints allows heading to be anticipated (Billington et al.,
Actively driving through the slalom environment recruited cortical regions known to play a role in processing visual motion, spatial updating and somatosensory processing. The location of the cingulate gyrus activation for both SteerNr and SteerPv corresponds well with the region indentified as CSv (Wall and Smith,
Activation common to both steering conditions was found in the central area of the precuneus, which is part of the SPL. The central precuneus region has previously been shown to exhibit strong connections to multimodal inferior parietal lobes regions and dorsolateral prefrontal cortex in humans. This activation may reflect a role in updating the visual location of objects in both conditions, which could be somewhat independent of intended actions (Wolbers et al.,
When participants had access to future path information they were able to use this information in order to execute smoother steering trajectories, and we would expect such strategies to be reflected at the neural level. Indeed, the SteerPv condition elicited much more extensive activations in the SPL and IPS, as well as additional activations in the primary motor cortex (BA4), somatosensory cortex, occipital, occipitotemporal regions, and the cerebellum. The nature of activations in SPL is suggestive of the neural processes engaged during SteerPv compared to SteerNr. Activation during the SteerNr trials was more posterior, toward the parieto-occipital fissure. This region is thought to play a stronger role in visual processing, receiving inputs from areas of visual cortex, including MT, and having indirect connections through MT and MST to parietal regions such as VIP in the macaque (Colby et al.,
Participants were allowed free eye movement during our experiment and, in general, SPL activation can often reflect the planning of eye movements (Kan et al.,
During steering trials participants were continually receiving feedback on the efficacy of motor commands on the basis of efferent information. The SteerPv trials required participants to continually adjust errors in heading in a manner that was not only appropriate to negotiate the immediate obstacle (as in SteerNr), but also optimal for traversing smoothly between the immediate obstacle and the next obstacle in the sequence based on their relative positions. Correction of ongoing movement using feedback loops is thought to recruit the IPS (Desmurget et al.,
The cerebellum, thalamus, middle, and inferior frontal gyri (IFG) and IPS, and occipitotemporal cortex have also been implicated in playing a role in visual feedback control of ongoing motor movement (Inoue et al.,
A final question posed in this study was regarding whether individual differences in steering performance could be identified in specific cortical regions. We found two regions in the right superior postcentral sulcus in which greater activation was predicted by smoother steering performance. This somatosensory cortical area has a similar locality to that deemed to be putative human VIP (Sereno and Huang,
This study has confirmed that a network of cortical areas play a role in effective locomotion by means of effective visuo-spatial encoding, visuo-motor encoding and integration and online corrective feedback mechanisms. In particular, the capability of neurons in regions of the SPL may allow for effective encoding of visual information regarding egocentric heading and obstacle location for the purpose of generating motor commands during locomotion. The importance of integrating visual and motor coordinates during locomotion is reinforced by our finding that participants who displayed smoother steering patterns showed greater activation in a region of the postcentral sulcus known to encode information from different sensory modalities. The IPS and cerebellum are engaged in order to estimate effector trajectories and compute error signals for correcting ongoing movements to support skilled motor actions. To spline a path through immediate and future waypoints multiple object goals have to be integrated and smoother steering is more appropriate. This paper implicates both dorsal steam processes and a parietal-cerebella network in not only supporting steering behaviors, but also contributing toward improved performance on an individual level.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
This work was supported by an award from the UK Engineering and Physical Sciences Research Council (EP/ D055342/1) to John P. Wann, Richard M. Wilkie, and David T. Field. We thank David T. Field, University of Reading, for his input into this study.