Mechanisms of directed attention in the human extrastriate cortex as revealed by functional MRI

Science

Volumn 282, Issue 5386, 1998, Pages 108-111

  Kastner, Sabine ^a   De Weerd, Peter ^a   Desimone, Robert ^a   Ungerleider, Leslie G ^a  

^a NATIONAL INSTITUTE OF MENTAL HEALTH   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ARTICLE; ATTENTION; BRAIN CORTEX; HUMAN; HUMAN TISSUE; NUCLEAR MAGNETIC RESONANCE IMAGING; PRIORITY JOURNAL; SENSORY SYSTEM; VISUAL FIELD; VISUAL STIMULATION; VISUAL SYSTEM;

ADULT; ATTENTION; BRAIN MAPPING; FIXATION, OCULAR; HUMANS; MAGNETIC RESONANCE IMAGING; MALE; NEURONS; PHOTIC STIMULATION; VISUAL CORTEX; VISUAL FIELDS; VISUAL PATHWAYS; VISUAL PERCEPTION;

EID: 0032475917     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.282.5386.108     Document Type: Article

References (38)

1. J. Duncan, Psychol. Rev. 87, 272 (1980); _ and G. Humphreys, ibid. 96, 433 (1989); R. Desimone and J. Duncan, Annu. Rev. Neurosci. 18, 193 (1995); G. Rees, C. D. Frith, N. Lavie, Science 278, 1616 (1997).
2. J. Duncan, Psychol. Rev. 87, 272 (1980); _ and G. Humphreys, ibid. 96, 433 (1989); R. Desimone and J. Duncan, Annu. Rev. Neurosci. 18, 193 (1995); G. Rees, C. D. Frith, N. Lavie, Science 278, 1616 (1997).
3. J. Duncan, Psychol. Rev. 87, 272 (1980); _ and G. Humphreys, ibid. 96, 433 (1989); R. Desimone and J. Duncan, Annu. Rev. Neurosci. 18, 193 (1995); G. Rees, C. D. Frith, N. Lavie, Science 278, 1616 (1997).
4. J. Duncan, Psychol. Rev. 87, 272 (1980); _ and G. Humphreys, ibid. 96, 433 (1989); R. Desimone and J. Duncan, Annu. Rev. Neurosci. 18, 193 (1995); G. Rees, C. D. Frith, N. Lavie, Science 278, 1616 (1997).
5. M. I. Posner, Quart. J. Exp. Psychol. 32, 3 (1980); A. Treisman, Comput. Vision Graphics Image Process. 31, 156 (1985); J. Driver and G. C. Baylis, J. Exp. Psychology 15, 448 (1989).
6. M. I. Posner, Quart. J. Exp. Psychol. 32, 3 (1980); A. Treisman, Comput. Vision Graphics Image Process. 31, 156 (1985); J. Driver and G. C. Baylis, J. Exp. Psychology 15, 448 (1989).
7. M. I. Posner, Quart. J. Exp. Psychol. 32, 3 (1980); A. Treisman, Comput. Vision Graphics Image Process. 31, 156 (1985); J. Driver and G. C. Baylis, J. Exp. Psychology 15, 448 (1989).
8. J. Reynolds, L. Chelazzi, S. J. Luck, R. Desimone, Soc. Neurosci. Abstr. 21, 1759 (1995); J. Moran and R. Desimone, Science 229, 782 (1985); S. Treue and J. H. R. Maunsell, Nature 382, 539 (1996); C. E. Connor, D. C. Preddie, J. L. Gallant, D. C. Van Essen, J. Neurosci. 17, 3201 (1997); E. K. Miller, P. M. Gochin, C. G. Gross, Brain Res. 616, 25 (1993).
9. J. Reynolds, L. Chelazzi, S. J. Luck, R. Desimone, Soc. Neurosci. Abstr. 21, 1759 (1995); J. Moran and R. Desimone, Science 229, 782 (1985); S. Treue and J. H. R. Maunsell, Nature 382, 539 (1996); C. E. Connor, D. C. Preddie, J. L. Gallant, D. C. Van Essen, J. Neurosci. 17, 3201 (1997); E. K. Miller, P. M. Gochin, C. G. Gross, Brain Res. 616, 25 (1993).
10. J. Reynolds, L. Chelazzi, S. J. Luck, R. Desimone, Soc. Neurosci. Abstr. 21, 1759 (1995); J. Moran and R. Desimone, Science 229, 782 (1985); S. Treue and J. H. R. Maunsell, Nature 382, 539 (1996); C. E. Connor, D. C. Preddie, J. L. Gallant, D. C. Van Essen, J. Neurosci. 17, 3201 (1997); E. K. Miller, P. M. Gochin, C. G. Gross, Brain Res. 616, 25 (1993).
11. J. Reynolds, L. Chelazzi, S. J. Luck, R. Desimone, Soc. Neurosci. Abstr. 21, 1759 (1995); J. Moran and R. Desimone, Science 229, 782 (1985); S. Treue and J. H. R. Maunsell, Nature 382, 539 (1996); C. E. Connor, D. C. Preddie, J. L. Gallant, D. C. Van Essen, J. Neurosci. 17, 3201 (1997); E. K. Miller, P. M. Gochin, C. G. Gross, Brain Res. 616, 25 (1993).
12. J. Reynolds, L. Chelazzi, S. J. Luck, R. Desimone, Soc. Neurosci. Abstr. 21, 1759 (1995); J. Moran and R. Desimone, Science 229, 782 (1985); S. Treue and J. H. R. Maunsell, Nature 382, 539 (1996); C. E. Connor, D. C. Preddie, J. L. Gallant, D. C. Van Essen, J. Neurosci. 17, 3201 (1997); E. K. Miller, P. M. Gochin, C. G. Gross, Brain Res. 616, 25 (1993).
13. J. S. J. Luck, L. Chelazzi, S. A. Hillyard, and R. Desimone [J. Neurophysiol. 77, 24 (1997)] reported that both suppressive interactions and attentional effects were larger in V4 when competing stimuli were presented simultaneously than when presented sequentially at different locations. It is not known whether the suppressive effects are due to inhibition or reduction of excitation.
14. Fourteen contiguous, coronal, 5-mm-thick slices were acquired in 12 to 16 series of 60 images each, starting from the posterior pole (in-plane resolution, 2.5 mm × 2.5 mm). Gradient echo, echo planar imaging was used [reception time (TR) = 3 s, echo time (TE) = 40 ms, flip angle = 90°] on a 1.5 Tesla GE magnet. Functional images were coaligned with a high-resolution anatomical scan taken in the same session (three-dimensional spoiled gradient echo sequence; TR = 15 ms; TE = 7 ms; flip angle = 30°; matrix, 256 × 256 voxels). Activations were identified by means of multiple regression analysis of the time series of MRI intensities in every voxel and two regressors of interest [K. J. Friston et al., Neuroimage 2, 45 (1995)], reflecting contrasts between (i) visual stimulation versus blank periods and (ii) sequential versus simultaneous presentations. Additional regressors were used to factor out variance due to between-run changes in mean intensity and within-run linear changes. The statistical significance (P < 0.05) of activated regions was assessed by an analysis based on the spatial extent of each region [K. Friston, K. J. Worsley, R. S. J. Frackowiak, J. C. Mazziotta, A. C. Evans, Hum. Brain Mapp. 1, 210 (1994)]. The fMRI time series, averaged over all activated voxels in a given region during visual stimulation versus blank presentations (thresholded at a Z score of 3.7) and over runs for each participant, are presented as group data. In experiment 2, time series analysis was performed only on those voxels activated in both the unattended and attended conditions. Statistical significance was assessed with repeated measures ANOVAs on the six (experiment 1) or five (experiment 2) peak intensities of the fMRI signal. For each subject, statistical maps and structural images were transformed into Talairach space [J. Talairach and P. Tournoux, Co-Planar Stereotactic Atlas of the Human Brain (Thieme, New York, 1988)] using the template from SPM96. Participants (four men, 22 to 34 years old) gave written informed consent for participation.
15. Fourteen contiguous, coronal, 5-mm-thick slices were acquired in 12 to 16 series of 60 images each, starting from the posterior pole (in-plane resolution, 2.5 mm × 2.5 mm). Gradient echo, echo planar imaging was used [reception time (TR) = 3 s, echo time (TE) = 40 ms, flip angle = 90°] on a 1.5 Tesla GE magnet. Functional images were coaligned with a high-resolution anatomical scan taken in the same session (three-dimensional spoiled gradient echo sequence; TR = 15 ms; TE = 7 ms; flip angle = 30°; matrix, 256 × 256 voxels). Activations were identified by means of multiple regression analysis of the time series of MRI intensities in every voxel and two regressors of interest [K. J. Friston et al., Neuroimage 2, 45 (1995)], reflecting contrasts between (i) visual stimulation versus blank periods and (ii) sequential versus simultaneous presentations. Additional regressors were used to factor out variance due to between-run changes in mean intensity and within-run linear changes. The statistical significance (P < 0.05) of activated regions was assessed by an analysis based on the spatial extent of each region [K. Friston, K. J. Worsley, R. S. J. Frackowiak, J. C. Mazziotta, A. C. Evans, Hum. Brain Mapp. 1, 210 (1994)]. The fMRI time series, averaged over all activated voxels in a given region during visual stimulation versus blank presentations (thresholded at a Z score of 3.7) and over runs for each participant, are presented as group data. In experiment 2, time series analysis was performed only on those voxels activated in both the unattended and attended conditions. Statistical significance was assessed with repeated measures ANOVAs on the six (experiment 1) or five (experiment 2) peak intensities of the fMRI signal. For each subject, statistical maps and structural images were transformed into Talairach space [J. Talairach and P. Tournoux, Co-Planar Stereotactic Atlas of the Human Brain (Thieme, New York, 1988)] using the template from SPM96. Participants (four men, 22 to 34 years old) gave written informed consent for participation.
16. Fourteen contiguous, coronal, 5-mm-thick slices were acquired in 12 to 16 series of 60 images each, starting from the posterior pole (in-plane resolution, 2.5 mm × 2.5 mm). Gradient echo, echo planar imaging was used [reception time (TR) = 3 s, echo time (TE) = 40 ms, flip angle = 90°] on a 1.5 Tesla GE magnet. Functional images were coaligned with a high-resolution anatomical scan taken in the same session (three-dimensional spoiled gradient echo sequence; TR = 15 ms; TE = 7 ms; flip angle = 30°; matrix, 256 × 256 voxels). Activations were identified by means of multiple regression analysis of the time series of MRI intensities in every voxel and two regressors of interest [K. J. Friston et al., Neuroimage 2, 45 (1995)], reflecting contrasts between (i) visual stimulation versus blank periods and (ii) sequential versus simultaneous presentations. Additional regressors were used to factor out variance due to between-run changes in mean intensity and within-run linear changes. The statistical significance (P < 0.05) of activated regions was assessed by an analysis based on the spatial extent of each region [K. Friston, K. J. Worsley, R. S. J. Frackowiak, J. C. Mazziotta, A. C. Evans, Hum. Brain Mapp. 1, 210 (1994)]. The fMRI time series, averaged over all activated voxels in a given region during visual stimulation versus blank presentations (thresholded at a Z score of 3.7) and over runs for each participant, are presented as group data. In experiment 2, time series analysis was performed only on those voxels activated in both the unattended and attended conditions. Statistical significance was assessed with repeated measures ANOVAs on the six (experiment 1) or five (experiment 2) peak intensities of the fMRI signal. For each subject, statistical maps and structural images were transformed into Talairach space [J. Talairach and P. Tournoux, Co-Planar Stereotactic Atlas of the Human Brain (Thieme, New York, 1988)] using the template from SPM96. Participants (four men, 22 to 34 years old) gave written informed consent for participation.
17. T's and L's (0.6° in size) were presented for 250 ms in random order and in different orientations at 4 Hz. The T/L task had a high attentional load [see also (12)], in order to ensure proper fixation and to prevent participants from covertly attending to the peripheral stimuli. Performance measured outside the scanner (75% correct on average) did not differ during blank, sequential, or simultaneous presentation periods [F(2, 143) = 1.60, P = 0.21]. Hence, neither presentation condition interfered with the T/L task, indicating that this task provided sufficient attentional load to preclude exogenous attentional cueing.
18. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
19. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
20. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
21. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
22. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
23. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
24. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
25. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
26. The borders of retinotopic areas in the ventral extrastriate cortex of humans and monkeys [R. Gattass, C. G. Gross, J. H. Sandell, J. Comp. Neurol. 21, 519 (1981); R. Gattass, A. P. Sousa, C. G. Gross, J. Neurosci. 8, 1831 (1988); M. I. Sereno et al., Science 268, 889 (1995); R. B. H. Tootell et al., J. Neurosci. 15, 3215 (1995); E. A. DeYoe et al., Proc. Natl. Acad. Sci. U.S.A. 93, 2382 (1996)] are formed by the representations of either the vertical (V1/V2 or VP/V4) or the horizontal (V2/VP) meridians. Meridians were mapped with color- and luminance-contrast checkered stimuli. In five of eight participants, it was difficult to determine the extent of VP, because the representations of the V2/VP and the VP/V4 border were abutting or overlapping [see S. Shipp, J. D. G. Watson, R. S. J. Frackowiak, S. Zeki, Neuroimage 2, 125 (1995)]. We will, therefore, refer to the area between the V1/V2 border and the VP/V4 border as "V2", although it likely contains parts of VP. The presumptive lower field representation of V4 was determined with the complex images presented to the lower field and was found to be located adjacent and lateral to V4's upper field representation on the fusiform gyrus [D. J. McKeefry and S. Zeki, Brain 120, 2229 (1997)]. The region we have termed V4 may include all or part of the region termed V8 by Hadjikhani et al. [Nature Neurosci. 1, 235 (1998)]. In the region located anterior to V4 (and also V8), the spatial segregation of upper and lower field representations was no longer seen, suggesting that this area was different from V4. Because area TEO is located just anterior to V4 in the monkey [D. Boussaoud, R. Desimone, L. G. Ungerleider, J. Comp. Neurol. 306, 554 (1991)], we will refer to this similarly located area as putative human TEO.
27. In a separate experiment, four stimuli (each 0.5° × 0.5° in size) were presented 6° apart from each other in the right upper quadrant. The prediction was that increasing the spatial separation between stimuli would strongly reduce suppressive interactions in areas with small (V2) and intermediate (V4) receptive fields but not in areas with large receptive fields (TEO) extending over a quadrant. Results from three participants showed that the interactions were indeed abolished in V2, were strongly reduced in V4, but were still present in TEO.
28. Three of the eight participants saw complex stimuli at 1 Hz in the following presentation configurations: one stimulus presented to the upper visual field, three presented to the lower visual field, or all four presented together. Participants performed the T/L task at fixation throughout the scan. All other presentation parameters were as in experiment 1.
29. The averaged signal changes in V4's upper field were 1.04% evoked by the single stimulus, 0.83% evoked by the four stimuli, and 0.52% evoked by the three stimuli in the lower field (due to signal spread into the upper field). Because of this spread, the actual suppression effect might be much larger than that reflected in the difference in responses to the single stimulus and to the four stimuli. The response differences were not significant in V1 and V2. Thus, with this experimental design, suppressive interactions could only be demonstrated in areas with sufficiently large receptive fields.
30. All four stimuli, including the stimulus selected to be the target, were randomly presented in all four locations in blocks of 15 s. The blocks with directed attention to the stimulus display were indicated by a marker presented close to the fixation point 1 s before the block started. In pilot experiments, we found that the attentional effect during the first attended block in a sequence was always stronger than in other attentional blocks within a run. To attenuate this attentional "onset" effect, each run started with a block of attended presentations that was discarded from analysis.
31. Before being scanned, participants received training in the directed attention task and fixation was monitored. During the directed attention task, targets were identified correctly at rates of 86 and 93%, respectively, in the sequential and simultaneous presentation conditions. The attentional load of the T/L task and the directed attention task was assessed by having participants perform them simultaneously in tests conducted outside the scanner. Both tasks interfered with each other when performed simultaneously. Performance in the directed attention task dropped significantly [F(1, 192) = 130.92, P < 0.0001] from 86 to 45% and from 93 to 49%, respectively, in the sequential and simultaneous conditions. Likewise, performance in the T/L task dropped significantly [F(1, 191) = 66.76, P < 0.0001] when participants were required to simultaneously identify targets at the target location. Thus, both tasks had a high attentional load. Participants rarely identified target stimuli in locations other than the attended location.
32. Because the cortical activations from the attended and unattended stimuli could not be separated, any increase in response to the attended stimulus might, in principle, be counterbalanced by a decrease in response to an unattended one, working against our hypothesis. However, the attended stimulus was located closest to the fovea and thus would dominate the response to the array because of the cortical magnification factor. Further, single-cell studies have shown that attention to a stimulus filters out the suppressive influence of nearby stimuli very effectively, but it has a smaller suppressive effect on the response to unattended ones (R. Desimone, unpublished observations).
33. 3 in TEO, averaged over participants. In the attended condition, brain volumes increased significantly in V4 and TEO but not in V1 and V2 [V4: 78 ± 16% (mean ± SEM); TEO: 120 ± 36%; ANOVA, main attentional effect: F(1, 64) = 14.2, P < 0.001; cortical area and attentional effect: F(3, 64) = 2.82, P < 0.05].
34. H. J. Heinze et al., Nature 372, 543 (1994); G. R. Mangun, Psychophysiology 32, 4 (1995); G. Rees, R. Frackowiak, C. Frith, Science 275, 835 (1997); R. Vandenberghe et al., J. Neurosci. 17, 3739 (1997).
35. H. J. Heinze et al., Nature 372, 543 (1994); G. R. Mangun, Psychophysiology 32, 4 (1995); G. Rees, R. Frackowiak, C. Frith, Science 275, 835 (1997); R. Vandenberghe et al., J. Neurosci. 17, 3739 (1997).
36. H. J. Heinze et al., Nature 372, 543 (1994); G. R. Mangun, Psychophysiology 32, 4 (1995); G. Rees, R. Frackowiak, C. Frith, Science 275, 835 (1997); R. Vandenberghe et al., J. Neurosci. 17, 3739 (1997).
37. H. J. Heinze et al., Nature 372, 543 (1994); G. R. Mangun, Psychophysiology 32, 4 (1995); G. Rees, R. Frackowiak, C. Frith, Science 275, 835 (1997); R. Vandenberghe et al., J. Neurosci. 17, 3739 (1997).
38. We thank J. M. Maisog, M. I. Elizondo, and M. A. Georgopoulos for help with data analysis; P. Jezzard for help with scanning; and J. V. Haxby, B. Jagadeesh, A. Martin, J. Reynolds, and U. Ziemann for valuable discussions. S.K. was supported by the Deutsche Forschungsgemeinschaft.