- Email: [email protected]
Position discrimination in one - versus two - dimensions: an fMRI-study
Neurolmage 11, Number 5, 2000, Part 2 of 2 Parts | 1 1 I~ ~ t
®
PERCEPTION
Position discrimination
in one - versus two - dimensions:
an fMRI-study
Gereon R. Fink*t, John C. Marshall,, Peter H. Weisst, Nadim J. Shaht, Ivan Tonit, Peter W. Halligan§, Karl Zillest ~
*Neurologische Klinik, Heinrich-Heine-Universitiit Diisseldo~ Germany ~flnstitut fiir Medizin, Forschungszentrum Jiilich, Germany SNeuropsychology Unit, The Radcliffe Infirmary, Oxford, UK §School of Psychology, Cardiff University, Cardiff, UK ~C. & O. Vogt- Hirnforschungsinstitut, Heinrich-Heine-Universitiit Diisseldo~ Germany Introduction Line bisection is widely used as a clinical test of spatial cognition in patients with left visuospatial neglect after right hemisphere lesion (1). Surprisingly, many neglect patients who show severe impairment on marking the center of horizontal lines accurately mark the center of squares (2,3). That these patients with left neglect are also typically poor at judging whether lines are correctly prebisected implies that the deficit is perceptual, not motoric. These findings suggest a differential neural basis for one- and two-dimensional visual position discrimination.
Subjects and Methods Normal subjects (n= 12) were scanned using fMRI. In C1 & C3 subjects indicated by right hand button press whether or not the square (CI: Squaremark task, SM) was correctly pre-marked in its center, or whether or not the line (C3: Landmark task, LM) was correctly pre-bisected. In C2 & C4 subjects indicated whether or not there was a mark anywhere within the square (C2: control for Squaremark task, SC) or anywhere on the line (C4: control for Landmark task, LC). A low level visual baseline (in which no visuospatial task was performed) served to establish the neural basis common to all four conditions. The factorial design allowed us to assess the neural activations associated with the different stimuli, with both positional judgement tasks, as well as the simple main effects of LM and SM. Differential activations observed during LM and SM were then tested for significance by assessing the interactions. The different visual inputs inherent to the SM and LM are thus controlled for. We emphasize that the design of this study does indeed isolate the operation of position discrimination per se after other task demands have been removed by subtraction of the control conditions. Functional MR images were acquired on a Siemens Vision 1.5T whole-body scanner using standard EPI imaging procedures and statistical analysis (using SPM 97; analyses included a random effects model) as described elsewhere (4).
Results Increases in neural activity (p<0.05, corrected) common to all four conditions (relative to baseline) were observed in networks concerned with visual stimulation and analysis (complete occipital cortex and temporo-occipital transition), visuospatial attention (superior and inferior parietal lobules, and thalamus), decision making (lateral premotor and supplementary motor cortices), visuomotor coordination (cerebellum and striatum), and response execution (primary motor and somatosensory cortices). Increases in neural activity during L M + S M relative to L C + S C were observed only in lateral inferior occipital cortex bilaterally (p<0.05 corrected). Increases in neural activity associated with the SM relative to SC (p<0.005) were observed in ventral posterior brain regions; no significant increases were observed in either the superior or inferior parietal cortex. As expected (4), LM relative to LC (p<0.005) activated lateral inferior and medial superior occipital cortex bilaterally, the left cerebellar hemisphere, and the right inferior parietal cortex. Analysis of the interaction terms confirmed the significance of the differential activations in the right intraparietal sulcus associated with LM (p<0.005, Z = 2.9) and in the lingual gyrus bilaterally associated with SM (p<0.005, Z = 3.2). There were no significant differences in eye movements or error rates across conditions. RTs revealed an advantage [F(1,40) = 7.15, p = 0.01] for SC (479 --- 31 ms) and LC (485 ± 39 ms) relative to LM (552 ± 90 ms) and SM (514 ± 72 ms).
Condusion In a previous fMRI study we demonstrated the normal functional neuroanatomy of line center judgements (4). The current data replicate these findings for line center judgements. Furthermore, the distinct neural bases observed for one- and two-dimensional visuospatial judgements help explain the previously observed clinical dissociations by showing that as a stimulus becomes a better, more 'object-like' gestalt, the ventral visuoperceptive route assumes more responsibility for assessing position within the object.
References (1) (2) (3) (4)
Heilman KM et al., In: Feinberg TE et al., eds. Behavioral Neurology and Neuropsychology, pp. 309-317. Halligan PW & Marshall JC (1991) Neuropsychologia 29:619-628. Tegner R, Levander M (1991) J Neurol Neurosurg Psychiatr 54:882-887. Fink GR, et al. (2000) Neurology: in press.
$687
®
PERCEPTION
Position discrimination
in one - versus two - dimensions:
an fMRI-study
Gereon R. Fink*t, John C. Marshall,, Peter H. Weisst, Nadim J. Shaht, Ivan Tonit, Peter W. Halligan§, Karl Zillest ~
*Neurologische Klinik, Heinrich-Heine-Universitiit Diisseldo~ Germany ~flnstitut fiir Medizin, Forschungszentrum Jiilich, Germany SNeuropsychology Unit, The Radcliffe Infirmary, Oxford, UK §School of Psychology, Cardiff University, Cardiff, UK ~C. & O. Vogt- Hirnforschungsinstitut, Heinrich-Heine-Universitiit Diisseldo~ Germany Introduction Line bisection is widely used as a clinical test of spatial cognition in patients with left visuospatial neglect after right hemisphere lesion (1). Surprisingly, many neglect patients who show severe impairment on marking the center of horizontal lines accurately mark the center of squares (2,3). That these patients with left neglect are also typically poor at judging whether lines are correctly prebisected implies that the deficit is perceptual, not motoric. These findings suggest a differential neural basis for one- and two-dimensional visual position discrimination.
Subjects and Methods Normal subjects (n= 12) were scanned using fMRI. In C1 & C3 subjects indicated by right hand button press whether or not the square (CI: Squaremark task, SM) was correctly pre-marked in its center, or whether or not the line (C3: Landmark task, LM) was correctly pre-bisected. In C2 & C4 subjects indicated whether or not there was a mark anywhere within the square (C2: control for Squaremark task, SC) or anywhere on the line (C4: control for Landmark task, LC). A low level visual baseline (in which no visuospatial task was performed) served to establish the neural basis common to all four conditions. The factorial design allowed us to assess the neural activations associated with the different stimuli, with both positional judgement tasks, as well as the simple main effects of LM and SM. Differential activations observed during LM and SM were then tested for significance by assessing the interactions. The different visual inputs inherent to the SM and LM are thus controlled for. We emphasize that the design of this study does indeed isolate the operation of position discrimination per se after other task demands have been removed by subtraction of the control conditions. Functional MR images were acquired on a Siemens Vision 1.5T whole-body scanner using standard EPI imaging procedures and statistical analysis (using SPM 97; analyses included a random effects model) as described elsewhere (4).
Results Increases in neural activity (p<0.05, corrected) common to all four conditions (relative to baseline) were observed in networks concerned with visual stimulation and analysis (complete occipital cortex and temporo-occipital transition), visuospatial attention (superior and inferior parietal lobules, and thalamus), decision making (lateral premotor and supplementary motor cortices), visuomotor coordination (cerebellum and striatum), and response execution (primary motor and somatosensory cortices). Increases in neural activity during L M + S M relative to L C + S C were observed only in lateral inferior occipital cortex bilaterally (p<0.05 corrected). Increases in neural activity associated with the SM relative to SC (p<0.005) were observed in ventral posterior brain regions; no significant increases were observed in either the superior or inferior parietal cortex. As expected (4), LM relative to LC (p<0.005) activated lateral inferior and medial superior occipital cortex bilaterally, the left cerebellar hemisphere, and the right inferior parietal cortex. Analysis of the interaction terms confirmed the significance of the differential activations in the right intraparietal sulcus associated with LM (p<0.005, Z = 2.9) and in the lingual gyrus bilaterally associated with SM (p<0.005, Z = 3.2). There were no significant differences in eye movements or error rates across conditions. RTs revealed an advantage [F(1,40) = 7.15, p = 0.01] for SC (479 --- 31 ms) and LC (485 ± 39 ms) relative to LM (552 ± 90 ms) and SM (514 ± 72 ms).
Condusion In a previous fMRI study we demonstrated the normal functional neuroanatomy of line center judgements (4). The current data replicate these findings for line center judgements. Furthermore, the distinct neural bases observed for one- and two-dimensional visuospatial judgements help explain the previously observed clinical dissociations by showing that as a stimulus becomes a better, more 'object-like' gestalt, the ventral visuoperceptive route assumes more responsibility for assessing position within the object.
References (1) (2) (3) (4)
Heilman KM et al., In: Feinberg TE et al., eds. Behavioral Neurology and Neuropsychology, pp. 309-317. Halligan PW & Marshall JC (1991) Neuropsychologia 29:619-628. Tegner R, Levander M (1991) J Neurol Neurosurg Psychiatr 54:882-887. Fink GR, et al. (2000) Neurology: in press.
$687