Selective Attention to Face Identity and Color Studied With fmri

Transcription

1 Human Brain Mapping 5: (1997) Selective Attention to Face Identity and Color Studied With fmri Vincent P. Clark, 1 * Raja Parasuraman, 2 Katrina Keil, 1 Rachel Kulansky, 1 Sean Fannon, 2 Jose Ma. Maisog, 1 Leslie G. Ungerleider, 1 and James V. Haxby 1 1 Laboratory of Psychology and Psychopathology, NIMH, NIH, Bethesda, Maryland Catholic University, Washington, D.C Abstract: Cortical areas associated with selective attention to the color and identity of faces were located using functional magnetic resonance imaging (fmri). Six subjects performed tasks which required selective attention to face identity or color similarity using the same color-washed face stimuli. Performance of the color attention task but not the face attention task was associated with a region of activity in the collateral sulcus and nearby regions of the lingual and fusiform gyri. Performance of both tasks was associated with a region of activity in ventral occipitotemporal cortex that was lateral to the color responsive area and had a greater spatial extent. These fmri results converge with results obtained from PET and ERP studies to demonstrate similar anatomical locations of functional areas for face and color processing across studies. Hum. Brain Mapping 5: , Wiley-Liss, Inc. Key words: brain mapping; color perception; face recognition; selective attention; extrastriate cortex; visual cortex; visual pathways; magnetic resonance imaging INTRODUCTION Functional brain imaging research has revealed two close links between selective attention and perceptual processing in the human brain. First, PET and ERP studies have shown that selective attention to a particular stimulus feature is associated with the selective modulation of neural activity in the same (or overlapping) brain regions that are active during the normal perception of that feature [Corbetta et al., 1991; Haxby et al., 1994, 1997; Clark and Hillyard, 1996]. Second, activity in unattended perceptual processing areas can Contract grant sponsor: NIH; Contract grant number: AG *Correspondence to: Dr. Vincent P. Clark, LBC, NIMH, NIH, 10/ 4C104, 10 Center Dr., MSC-1366, Bethesda, MD vc@alw.nih.gov Received for publication 20 March 1997; accepted 27 March 1997 be attenuated [Haxby et al., 1994; Clark et al., 1996]. The modulation of neural activity by selective attention to object and spatial features is consistent with the division of visual function into ventral and dorsal processing streams, respectively [Ungerleider and Mishkin, 1982]. We have investigated whether similar modulation can be observed within one of the processing streams, the ventral object vision pathway. Previous research has suggested that the processing of face identity and object color may be performed within adjacent or partially overlapping areas of ventral occipital and temporal cortex. This is suggested by the results of a number of PET, fmri, and ERP studies which have examined the topography of face or color processing [Corbetta et al., 1991; Zeki et al., 1991; Allison et al., 1994; Haxby et al., 1994; Puce et al., 1995; Sakai et al., 1995; Clark et al., 1996]. Previously, face and color processing has been investigated in separate 1997 Wiley-Liss, Inc. This article was prepared by a group consisting of both United States government employees and non-united States government employees, and as such is subject to 17 U.S.C. Sec. 105.

2 Clark et al. studies using different stimuli. No study has examined selective attention to color vs. face identity in the same subjects using the same physical stimuli, which allows for the examination of selective modulation of brain activity independent of stimulus-specific processing. In the present study, we used fmri to study the functional neuroanatomy of face and color processing with the aim of 1) identifying the cortical regions involved in face and color processing; 2) comparing these results with previous fmri, PET, and ERP studies of face and color processing; and 3) examining whether selective modulation of neural activity within ventral stream areas can be demonstrated as a function of feature-specific attention. MATERIALS AND METHODS Six neurologically normal right handed volunteers (three women), ages years participated in this study. Stimuli were generated by a Power-Macintosh computer (Apple, Cupertino, CA) using SuperLab (Cedrus, Wheaton, MD) and were projected with a magnetically shielded LCD video projector (Sharp, Mahwah, NJ) onto a translucent screen placed at the feet of the subject. The subject was able to see the screen by the use of a mirror system or prism glasses. The color and face attention tasks employed identical stimuli, requiring selective attention to color similarity or face identity during different experimental runs (Fig. 1A). Each item began with the presentation of one sample stimulus (1.5 sec duration) positioned centrally, followed by the presentation of two test stimuli (2.0 sec duration) positioned laterally, with 250 msec between successive items and between successive trials. In the face attention task, subjects indicated whether the right or left choice stimulus pictured the same individual as the sample stimulus by responding with buttons placed in their right and left hands, respectively. In the color attention task, subjects indicated which choice stimulus had a color that was closest along the color spectrum to the color of the sample stimulus. Stimulus parameters were chosen so that perceptual discriminations based on color and face identity were of equivalent difficulty, as measured by RT. The sensorimotor control task used monochrome scrambled face stimuli, which were presented in the same locations and rate as the attention stimuli. Subjects responded to the presentation of each double control stimulus by pressing both buttons. Subjects were pretrained on the tasks before MRI scanning. For three subjects (S1 S3), four alternating blocks of control and attention tasks, 21 s long for each task, were performed for each run. For S4 S6, three alternating 30 s long blocks of control and attention tasks were performed for each run. Face and color attention tasks were performed in separate runs for all subjects. Each task was preceded by a 1 s warning stimulus. All imaging used a 1.5 T MRI scanner (GE Signa, Milwaukee, WI) with gradient head coils designed for echo-planar imaging. A standard 5 surface radio frequency coil positioned beneath the occiput was used for three subjects (S1 S3), and a local gradient coil (Medical Advances, Milwaukee, WI), was used for the remaining three subjects (S4 S6). An interleaved multislice gradient echo EPI scanning sequence was used with a 64 X 64 matrix (echo time (TE) 40 ms, flip angle 90 ). In the three subjects studied with the surface coil (S1 S3), the location of the imaging volume was positioned beginning at the occipital pole, with 12 coronal sections, each 5 mm thick, a 16 cm field of view, and a repetition time (TR) of 3 s. In the three subjects studied with the whole-head RF coil (S4 S6), the same EPI sequence was used with 3.75 mm thick sagittal sections with isotropic voxels, which covered all of the head in 38 slices with a TR of 6 s, with a 24 cm field of view. For all studies, high-resolution structural Figure 1. A: A sample stimulus item used in the face and color attention tasks is shown at top, and a test stimulus item is shown below this. The sensorimotor control stimuli are shown below the stimuli for the attention tasks. B: Coronal sections from four subjects which pass through an area of significant interaction of color and face attention with control (shown in yellow) located near the collateral sulcus. Areas showing significant activation during both the face and color tasks (shown in orange) were located more laterally, near the fusiform gyrus and lateral occipitotemporal sulcus. The right of the brain is displayed on the left of the image. C: Signal amplitude averaged over repeated cycles of control (-30 to 0 s), attention (0 to 30 s), and control (30 to 60 s) tasks during runs of color attention (plotted in yellow) and face attention (plotted in red) for a region of interest located in the collateral sulcus that revealed a significant response during color attention but not during face attention of the subject shown in Figure 1D. Percentage change of signal intensity is shown relative to the first control baseline. D: ROIs for main effect of attention vs. control tasks (shown in orange) and for interaction of color and face attention with control (shown in yellow) from one subject plotted on an oblique section passing through the inferior cortical surface re-constructed from sagittal sections. E: Areas found by Allison et al. [1994] using implanted electrodes that were responsive to face presentations (shown in red) and to a color adaptation task (shown in yellow). F: Results from a series of PET studies of face and color attention. Areas shown in red were found by Haxby et al. [1994] to be active during a face matching task. Foci shown in yellow were found by Corbetta et al. [1991] to be active during color attention. Foci shown in blue were found by Zeki et al. [1991] to be active during color perception. 294

3 Selective Attention to Face Identity and Color Figure

4 Clark et al. images were also acquired at the same locations as the echo-planar images. These provided detailed anatomical information for later analysis. Significant changes in signal intensity for scans acquired during the face or color attention tasks relative to those acquired during the control task were taken to indicate changes in brain activity that were associated with the face and color processing operations required to perform these tasks. The statistical significance of differences in signal intensity was evaluated in each individual voxel using a repeatedmeasures analysis of variance (ANOVA). Areas where the changes in signal intensity between the attention and control tasks were significantly different for face attention runs relative to color attention runs indicated voxels where a significant effect of attention occurred. These areas were identified by an analysis of the interaction term of face attention vs. its control by color attention vs. its control. Statistical analyses were restricted to brain voxels with adequate signal intensity by selecting voxels with an average intensity of at least 25% of the maximum value across voxels. Between scan movement was corrected with Automatic Image Registration (AIR) software [Woods et al., 1993]. Groups of contiguous voxels significant at P.01 were used to define regions of interest (ROI) that were subsequently analyzed for mean location and volume and were used to acquire time series data averaged across their voxels. Average time series were computed across all voxels in each ROI and analyzed for the amplitude, latency, and time course of response during the face and color attention tasks and the control task. The mean locations of these ROIs in the Talairach coordinate system [Talairach and Tournoux, 1988] were found by computing the scaled distance between the mean location of the ROI in each subject and the maximum medial, lateral, dorsal, and ventral extent of the brain in that subject s coronal section positioned close to that ROI. Scaling along the y axis (anterior-posterior) could not be determined in the 12 slice data sets as the frontal pole was not imaged. Instead, this was estimated by comparing the distance from the occipital pole with the Talairach Atlas, without scaling for the maximum extent of cortex along this dimension. RESULTS AND DISCUSSION An area of increased signal was evoked near the right collateral sulcus and lingual gyrus during the color attention task but not during the face attention task in four subjects (Fig. 1B). Of the other two subjects, one (S5) had a significantly longer mean reaction time for the face attention task relative to the color attention task. Therefore, this subject was excluded from further analysis of selective attention as these two tasks were not of equivalent difficulty. The sixth subject had substantial movement artifacts which could not be corrected using AIR, and the data could not be analyzed for discrete areas of activity. This color attention region was located an average of 34 mm anterior to the occipital pole, and had a mean volume of 0.65 cc (SD 0.45 cc) and a mean Z score value of 3.1 (SD 0.69). The time series for this area showed an increased response during the color attention task when compared to the face attention task, as shown in Figure 1C for one subject with its location shown in Figure 1D and the lower right section of Figure 1B. While there was also a small response during the face attention task in this region, it was not statistically significant. Behavioral performance for the color and face attention tasks was equivalent, indicating that differences in patterns of cortical activation were not due to differences in task difficulty. Average reaction time on the face attention task (886 msec, SD 98 msec) was nearly equal to the average reaction time on the color attention task (890 msec, SD 107 msec). Hit rate was near 100% for both conditions. The mean location of this color attention region in Talairach coordinates [Talairach and Tournoux, 1988] was x 19 (SD 9.6), y -69 (SD 4.6), z -6.7 (SD 5.6). This location is within 1 cm of a region identified in a PET color attention task by Corbetta et al. [1991] (x 21, y -73, z 2), within 5 mm of a region identified in a PET color perception task by Zeki et al. [1991] (x 20, y -66, z -4) (Fig. 1F), and within 7 mm of a fmri color perception task by Sakai et al. [1995] (x 21, y -63, z -10). It is also close to areas identified using electrocorticographic recordings during color adaptation studies by Allison et al. [1994] (Fig. 1E). Many regions were found that were active during both the face and color attention tasks relative to the sensorimotor control task. Signal increases evoked during both attention tasks were found bilaterally in all subjects in ventral occipitotemporal cortex, extending forward from ventrolateral occipital cortex in the inferior occipital sulcus to the lateral occipitotemporal sulcus and fusiform gyrus, as shown in Figure 1D for one subject using the head coil. Additional areas of increased signal were found in the intraparietal sulcus and adjacent regions of the superior parietal gyrus, and in areas of cingulate and frontal cortex. These areas were also found to be active during previous PET and fmri studies of face perception using simulta- 296

5 Selective Attention to Face Identity and Color neously presented faces [Haxby et al., 1994; Clark et al., 1996]. Some regions of ventral occipitotemporal cortex were also found in each subject that responded significantly more during the face task than during the color task, but the precise locations of these regions were highly variable across subjects, making their significance unclear. While these results appear to be discordant with previous findings of a face-specific processing area that is located anteriorly and laterally to a color-specific processing area in the collateral sulcus [Allison et al., 1994; Haxby et al., 1994; Puce et al., 1995], a number of explanations can be offered for these results. Subjects may have processed face identity during both tasks, even though this information was irrelevant for the color attention task. This might be due to the fact that faces are visual objects that elicit obligatory responses, which are not significantly enhanced with attention to face identity. Alternatively, many of the neural fields involved in the color attention task may be interspersed with neurons that are active during face attention. If substantial numbers of neurons involved in each task were both positioned within the same voxel location, then activity would be registered in that voxel for both tasks. A further possibility is that the use of monochromatic colored stimuli for the attention tasks, when compared with the response to the monochromatic gray stimuli used in the control task, generated a response that was similar for both tasks. However, this is unlikely to be the case, as the regions previously found to be sensitive to the perception of color [Zeki et al., 1991; Sakai et al., 1995] are located much closer to the color attention region of the collateral sulcus than to regions found to respond equally in both tasks in the present study. CONCLUSION Attention to color activated a region of the right collateral sulcus close to regions found in previous studies to be activated by the passive response to colored stimuli and by selective attention to color. These results support the hypothesis that selective attention to object features involves the selective modulation of activity in the same areas that process those features. The results also demonstrate a high degree of correspondence between the results of PET, ERP, and fmri studies of face and color perception. ACKNOWLEDGMENTS We are very grateful to Drs. Avi Karni, Peter Jezzard, and the staff of the NIH NMR Research Center for their assistance. Dr. Roger Woods graciously provided us with the AIR software for image registration. R.P. was supported in part by NIH grant AG REFERENCES Allison T, McCarthy G, Nobre A, Puce A, Belger A (1994): Human extrastriate visual cortex and the perception of faces, words, numbers, and colors. Cerebral Cortex 4: Clark VP, Keil K, Maisog JM, Courtney SM, Ungerleider LG, Haxby JV (1996): Functional magnetic resonance imaging of human visual cortex during face matching: A comparison with positron emission tomography. NeuroImage 4(1):1 15. Clark VP, Hillyard SA (1996): Spatial selective attention affects early extrastriate but not striate components of the visual evoked potential. J Cogn Neurosci 8: Corbetta M, Miezin FM, Dobmeyer S, Shulman GL, Petersen SE (1991): Selective and divided attention during visual discriminations of shape, color, and speed; Functional anatomy by positron emission tomography. J Neurosci 11: Haxby JV, Clark VP, Courtney SM (1997): Distributed hierarchical neural systems for visual memory in human cortex. In: Hyman B, Duyckaerts C, Christen Y (eds): Connections, Cognition, and Alzheimer s Disease. Berlin: Springer: pp Haxby JV, Horwitz B, Ungerleider LG, Maisog JM, Pietrini P, Grady CL (1994): The functional organization of human extrastriate cortex: A PET-rCBF study of selective attention to faces and locations. J Neurosci 14: Puce A, Allison T, Gore JC, McCarthy G (1995): Face-sensitive regions in human extrastriate cortex studied by functional MRI. J Neurophysiol 74: Sakai K, Watanabe E, Onodera Y, Uchida I, Kato H, Yamamoto E, Koizumi H, Miyashita Y (1995): Functional mapping of the human colour centre with echo-planar magnetic resonance imaging. Proceedings of the Royal Society of London. Series B: Biological Sciences 261: Talairach J, Tournoux P (1988): Co-Planar Stereotaxic Atlas of the Human Brain. New York: Thieme Medical Publishers. Ungerleider LG, Mishkin M (1982): Two cortical visual systems. In: Ingle DJ, Goodale MA, Mansfield RJW (eds): Analysis of Visual Behavior. Cambridge, MA: MIT Press, pp Woods RP, Mazziotta JC, Cherry SR (1993): MRI-PET registration with an automated algorithm. J Comp Assist Tomo 17: Zeki S, Watson JD, Lueck CJ, Friston KJ, Kennard C, Frackowiak RS (1991): A direct demonstration of functional specialization in human visual cortex. J Neurosci 11: