Transient activation of inferior prefrontal cortex during cognitive set

© 1998 Nature America Inc. • http://neurosci.nature.com

article

Transient activation of inferior prefrontal cortex during cognitive set shifting


Seiki Konishi1, Kyoichi Nakajima1, Idai Uchida1, Masashi Kameyama1, Kiyoshi Nakahara2, Kensuke Sekihara2, and Yasushi Miyashita1,2,3 1

Department of Physiology, University of Tokyo School of Medicine, Hongo, Tokyo 113, Japan

2

Japan Science and Technology Corporation, Yushima, Tokyo 113, Japan

3

National Institute for Physiological Sciences, Okazaki 444, Japan Correspondence should be addressed to S.K. ([email protected])

The Wisconsin Card Sorting Test, which probes the ability to shift attention from one category of stimulus attributes to another (shifting cognitive sets), is the most common paradigm used to detect human frontal lobe pathology. However, the exact relationship of this card test to prefrontal function and the precise anatomical localization of the cognitive shifts involved are controversial. By isolating shift-related signals using the temporal resolution of functional magnetic resonance imaging, we reproducibly found transient activation of the posterior part of the bilateral inferior frontal sulci. This activation was larger as the number of dimensions (relevant stimulus attributes that had to be recognized) were increased. These results suggest that the inferior frontal areas play an essential role in the flexible shifting of cognitive sets.

The prefrontal cortex enables us to adapt to changing environments by permitting shifts from one mental state (cognitive set), directed toward a particular reaction tendency, to another. The Wisconsin Card Sorting Test (WCST)1,2 is the most common paradigm used to detect human frontal lobe pathology3–6. In this test (Fig. 1a), subjects are presented with a target card and four reference cards that may or may not be matched to the target card with respect to various stimulus dimensions, such as color, form or number of objects. Only one of these dimensions is relevant in determining the correct match; subjects must identify the matching card, learning by trial and error which is the relevant dimension to which they must attend. The relevant

dimension is intermittently changed, and subjects must therefore shift their cognitive set to identify and attend to the new dimension. Performance in this test is impaired by damage to the dorsolateral prefrontal cortex in both humans3 and monkeys7,8, which is thought to reflect a loss of ability to shift cognitive sets. However, the prefrontal cortex performs many different functions9–15, and it remains controversial whether the effects of prefrontal damage reflect specific loss of set shifting; moreover, the exact anatomical location of the set shifting function remains unknown. Because this shift-related neuronal activity should be transient by nature, we isolated the transient shift-related signals by

Table 1 Areas showing statistically significant activity Areas of activation (Brodmann area) R inferior frontal sulcus (BA 45/44) L inferior frontal sulcus (BA 45/44) R supramarginal gyrus (BA 40) L supramarginal gyrus (BA 40) Anterior cingulate gyrus (BA 24/32)

Coordinates (x, y, z)

S1

S2

Detection criteria and peak t-value S3 S4 S5

(39, 15, 22)

both (8.1)

both (6.8)

both (7.5)

both (7.9)

(-40, 18, 22)

both (8.6)

repl (6.0)

repl (5.3)

(38, -33, 39)

-

both (5.9)

(-45, -36, 48)

-

(-3, 25, 35)

thres (5.1)

S6

S7

both (6.8)

both (7.2)

-

both (8.3)

thres (4.8)

both (9.7)

both (11.1)

repl (5.6)

-

-

both (6.8)

-

repl (7.8)

repl (7.0)

both (7.0)

-

-

thres (7.0)

-

both (6.8)

thres (5.2)

-

-

-

The areas showing significant activity in at least three out of seven subjects (S1-S7) are listed. repl, significant by replication approach; thres, significant by thresholding approach; both, significant by both approaches.

80

nature neuroscience • volume 1 no 1 • may 1998


article

a

Fig. 1. The Wisconsin Card Sorting Test (WCST) and eventrelated fMRI method. (a) The computerized WCST used in this study. A five-card (one target card and four reference cards) stimulus presented once per trial is shown at the left. Subjects were required to match the target card to one of the four reference cards based on one of three possible dimensions (color, form and number) by pushing a particular response button shown on the right. The dimension was shifted after at least 10 consecutive correct trials. (b) The data-processing strategy to isolate the shift-related signals. Each small downward-pointing arrow indicates one trial, and each large downward-pointing arrow indicates one dimensional shift. Because it is at the onset of the presentation of ‘incorrect’ feedback stimulus given immediately after the dimensional shift that the subjects first realized the dimensional shifts, we defined this moment as time zero, and compared the images taken at each time point after the dimensional shifts with the images taken before the shifts (see Methods).

Results Functional images were obtained every two seconds per slice while

Percentage

using a novel event-related functional magnetic resonance imaging (fMRI)16 method (Fig. 1b), which enables us to identify transiently activated areas using the temporal resolution of fMRI. We found reproducible transient shift-related MR signals in the posterior part of the bilateral inferior frontal sulci. The signals were larger as the number of relevant dimensions was increased (loaddependent increase), confirming that the signals reflected the shift-related neuronal activity. These observations suggest that the inferior frontal cortical areas play essential roles in flexible shifting of cognitive sets. In addition to the prea b frontal cortex, we found less reproducible signals in other brain regions. 3D This may explain why the specificity 3D of the WCST as a measure of frontal 2D 2D lobe dysfunction has often been doubted17. Furthermore, the finding of this localized activation pattern in the prefrontal cortex helps in the characterization of the set-shifting funcSeconds Trial tion and the understanding of the functional organization of the human Fig. 2. The normalized distribution of the number of trials (a) and the time (b) taken to comprefrontal cortex. plete set shifting in the 3D and 2D conditions. Completion was defined as three or more consecPercentage


b

seven normal subjects performed our computerized WCST. We designed three conditions of the test to alter cognitive load: a three-dimensional (3D) condition in which all the three dimensions were used similarly to the original WCST, a two-dimensional (2D) condition in which two of the three dimensions were used and alternated, and a one-dimensional (1D) condition, which did not contain shifts. The dimensions used in each run of the 2D and 1D conditions were counterbalanced among the three dimensions, and subjects were informed of the relevant dimensions prior to each run. In the 1D condition, the ‘incorrect’ feedback stimulus (visual presentation of an X) was presented regardless of the response, but set-shifting was not required. This served as a sensory control. From the performance data, we roughly estimated the duration of the shift-related neuronal activity (Fig. 2). The mean durations were 3.2 seconds (1.7 trials) in the 3D condition and 1.8 seconds (1.0 trial) in the 2D condition, and the difference between the durations of these conditions was significant (paired t-test, p< 0.0001). It is reported18–26 that short-duration neuronal activity elicits MR signals that peak 5 to 9 seconds after the neuronal activity. Therefore, the MR signals elicited by the short shift-related neuronal activity should be detected transiently within the time window of 5 to 9 seconds after the shift. The strategy used to isolate shift-related fMRI signals revealed a small number of focal regions that showed transient responses (5.9 regions per subject on average). Such activity was most reproducibly observed in the posterior part of the inferior frontal sulci of both hemispheres (Fig. 3a). In these regions, the transient signals peaked 7 seconds after the dimensional shifts in the 3D condition, consistent with the durations of the shift-related activity estimated above. We applied two different detection criteria (replication27 and thresholding28) to these regions, and the activity turned out to be significant by the criteria of both these methods. Both of these regions showed greater activation at 7 seconds after the shift in the 3D condition than in the 2D (onetailed t-test, p< 0.05 with Bonferroni correction) and 1D conditions (p< 0.005) (Fig. 3b). Figure 4a demonstrates the across-subject reproducibility of the shift-related activity from the pooled data of all the seven subjects. The two detection criteria (replication and thresholding) revealed highly reproducible activity in the right and left inferior frontal sulci. These activated regions (n = 6 for the right sulcus and n = 7 for the left sulcus) were located within the range of x = 39 ± 7 mm, y = 15 ± 5 mm, z = 22 ± 3 mm (right) and x =

utive correct responses. The time taken was defined as the interval between the onset of the ‘incorrect’ feedback stimulus after the first incorrect trial after the dimension change, and the onset of the ‘correct’ feedback stimulus after the first correct trial in a sequence of three correct trials. Error bars indicate the SD of the seven subjects.


81


a

Seconds

b

Left inferior frontal sulcus

Right inferior frontal sulcus Percentage

Percentage

Fig. 3. Shift-related activity for one subject detected by event-related fMRI method. (a) Bilateral activity in the inferior frontal sulci time-locked to the dimensional shifts in the 3D condition in one subject. The pixel-wise statistical significance level from 7 s after the shifts in the 3D condition is color-coded and mapped on the corresponding T1-weighted image. The activated regions in the right and left inferior frontal sulci are enlarged, and the activated pixels at 1s to 13 s after the shifts are shown sequentially in panels. The color scale at top left represents the p-value calculated from all the data in the 3D condition using paired t-test (for statistical evaluation, see data analysis). The yellow color at the bottom represents p = 0.0005, and the red color at the top represents p = 1 x 10–6. (b) The time courses of percent increase of the shift-related signals in the right and left inferior frontal sulci (shown in the left and right figures respectively) of the same subject in the 3D, 2D and 1D conditions. First, a region of interest (ROI) was delineated by the hypothesis-generating data set of the 3D condition (see Methods). Then the time course of the ROI was calculated from the hypothesis-testing data set of the 3D condition and the data of the 2D and 1D conditions.

Seconds

Seconds

-40 ± 8 mm, y = 18 ± 6 mm, z = 22 ± 3 mm (left) of Talairach’s coordinates29. Both of these regions showed greater activation at 7 seconds after the shift in the 3D condition than in either the 2D or the 1D condition (one-tailed t-test, p< 0.05 with Bonferroni correction) (Fig. 4b). This load-dependent increase suggests that the signals reflect the neural mechanisms mediating set-shifting. According to the atlas of Talairach and Tournoux, the probable Brodmann areas (BA) activated in the inferior frontal sulci are the dorsal parts of BA 45/44 (Fig. 5). However, we also found other activated regions in the right and left supramarginal gyri and the anterior cingulate cortex, although these regions were less reproducibly activated among the subjects (Table 1).

Percentage

Discussion The experimental strategy used here to isolate the transient shift-related MR signals revealed reproducible activation of a small number of foci in the posterior part of the right and left inferior frontal sulci. We observed no other areas that showed such reproducible activity. This is surprising because the shiftrelated signals may reflect several cognitive components that implement the set-shifting function, such as recognition of the incorrect feedback stimulus, selection of a new dimension by abstract reasoning, internal speech, response inhibition, and updating working memory. It is unlikely that the activity in the bilateral inferior frontal sulci reflected the recognition of the incorrect Right inferior frontal sulcus a b feedback stimulus because no activation was observed in the 1D condition (Fig. 4b). It is also unlikely that the activity in the left inferior frontal sulcus was derived from internal speech because the region is Seconds located 10 to 15 mm superior to Broca’s area 30,31 , Left inferior frontal sulcus where speech related activation is normally observed. Fig. 4. Across-subject distribution of shift-related activity. The regions of activation (a) Locations of all the shift-related regions identified by we observe are also separatthe two detection criteria (replication and thresholding) in ed by 15-25 mm from the the 3D condition of the seven subjects. The locations of mid-dorsal areas (BA46/49) the pixels that have maximal t-values within the regions are that have been reported to plotted using symbols of different colors: red, significant by be activated in previous the replication approach; green, significant by thresholding Seconds studies of working-memoapproach; yellow, significant by both approaches. They are ry tasks 32–43 . However, a mapped on eight corresponding anatomical slice images of a representative subject after matching the subjects’ anatomical images to the atlas of Talairach and part of the working memoTournoux using translation, rotation and stretching in two dimensions. (b) Similar to Fig. 3b but for ry system11,44 (for instance, the average from six subjects whose regions in the right or left inferior frontal sulcus showed signifi- updating the contents of cant activity by the replication approach. working memory) may Percentage


article

82



article

ness = 60 mm)49 and excluded from analysis regions that overlapped with detected veins.


Fig. 5. The centers of shift-related regions in the right and left inferior frontal sulci projected on to the lateral surface of 3D images (red spots).

involve the inferior frontal sulcus40,41. In addition, the activation in the right inferior frontal sulcus may reflect neural mechanisms underlying response inhibition function, as shown in our recent study45. Although understanding the precise nature of these activated regions requires further study, the localized activation seen here contributes to our understanding of the heterogenous functional organization of the prefrontal cortex9–15. Our results are consistent with previous positron emission tomography studies 46,47 reporting that the dorsolateral prefrontal cortex was activated to a greater degree during the performance of the WCST than during the performance of control tasks in which set-shifting was not required. The present study not only specified the activated regions within the dorsolateral prefrontal cortex, but also subtracted out shift-independent signals in the posterior cortices, such as regions in the occipital cortex that had been reported46,47 to be more strongly activated than the dorsolateral prefrontal cortex. We also observed transient shift-related signals in the parietal cortex, and although these were less reproducible than those in the prefrontal cortex, this suggests that the WCST may not be a completely specific indicator of frontal lobe dysfunction 17 . Nonetheless, the highly reproducible activity observed in the bilateral inferior frontal sulci suggests that these regions play an essential role in the ability to shift cognitive sets. Methods COGNITIVE TASK. In each trial of the test, the five-card stimulus shown in Fig. 1a was presented until next image scan, which occurred less than 0.5 s after the subject’s response, after which the second stimulus was presented. Therefore, the length of the stimulus presentation depended on the subjects’ reaction time, and varied by 0.5 s in a step-wise fashion paced by the image scan. After a 0.1 s delay from the image scan, a feedback stimulus (correct: O, incorrect: X) was presented for 0.5 s. The fivecard stimulus of the next trial was then presented 0.1 s after the end of the feedback stimulus. Because it took at least 0.5 s for the subjects to respond, one trial lasted at least 1.5 s. The target card stimuli were randomly selected from a pool of two decks of 64 cards each. The total number of shifts analyzed in each slice was 34 to 36 for 3D, 17 to 18 for 2D and 17 to 18 for 1D conditions.

IMAGE SCAN. Seven healthy right-handed48

male subjects participated in this study. We used 1.5 T gradient-echo echo-planar imaging sequences (TR = 2 s, TE = 20 ms, flip angle = 90°)20,45,49,50. The range of z = 6 mm to 54 mm at y = 0 mm (oblique by 10°) of Talairach’s coordinates was covered by eight contiguous transverse slices (slice thickness = 6 mm, in-plane resolution = 3 × 3 mm2, obtained with two sets of four slices). We took T1-weighted spin-echo images of the corresponding slices every six runs to estimate head movement and rejected runs in which head movement greater than 1.5 mm in any direction had occurred. We also performed magnetic resonance angiography scans (three-dimensional, phase-contrast sequence: TR = 40 ms, TE = 16 ms, flip angle = 20°, flow for a phase shift of π = 10 cm/s, slice thickness = 2.5 mm, bulk thicknature neuroscience • volume 1 no 1 • may 1998

DATA ANALYSIS. The shift-related fMRI signals were detected using the following procedure. The time zero was defined as the time at onset of the presentation of incorrect feedback stimulus given immediately after the dimensional shift, and all the image data of each slice were binned as 5, -3, -1, 1, 3, 5, 7, 9, 11 or 13 s pools for data analysis (Fig. 1b). Then we estimated the across-shift mean and variance of the difference between the images taken at each time point after time zero and the averaged images obtained from three time points before time zero. We set a time window of 5 to 9 s after the onset of the incorrect feedback stimulus to detect activated areas related to the set-shifting, because it is within this time window that we can effectively detect the transient signals elicited by the brief neuronal activity that begins at time zero18–26. In this study, two approaches were used to identify activated regions: a replication approach27 and a thresholding approach28. In the replication approach, data were divided into two separate sets by alternately classifying the data from each shift as the hypothesis-generating data set and the hypothesis-testing data set. The hypothesis-generating data set was used to delineate regions of interest (ROIs) with four or more contiguous pixels above the significance level of p< 0.005 (two-tailed paired t-test) within the time window. The hypothesis-testing data set was then used to test whether the ROIs replicated with a significance level of p< 0.005 (two-tailed paired t-test) within the time window of 5, 7 and 9 s (threefold Bonferroni-corrected). In the thresholding approach, we delineated activated regions with eight or more contiguous pixels above p< 0.0005 (two-tailed paired t-test) within the time window using the whole data sets. To show that the activated regions determined by these two significance criteria were truly positive, we conducted additional control experiments and estimated the number of false-positive regions that appeared where no experimental effects of shift-related activity were expected. In these experiments, the same subjects chose reference card stimuli without shifts throughout runs, and image data was collected and analyzed in the same way as for the 3D condition. When the replication approach was used, the number of false-positive regions averaged from the seven subjects in any of the eight slices scanned in this study was 0.016 per time point, and the estimated probability of one or more false-positive regions in three time points (i.e., the number of the critical time points from five to nine s) was less than 5% as determined by a Poisson distribution. In the thresholding approach, the criterion was more lenient than that in the replication approach, and the estimated probability of one or more false-positive regions within the prefrontal cortex in three time points was less than 5%.

Acknowledgments S. K. is supported by JSPS Research Fellowships for Young Scientists. This work was supported by a Grant-in-Aid for Specially Promoted Research from the Japanese Ministry of Education, Science and Culture to Y. M., and by a grant from Nissan Science Foundation to Y. M.

RECEIVED 6 FEBRUARY: ACCEPTED 11 MARCH 1998 1. Grant, D. A. & Berg, E. A. A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Weigl-type card-sorting problem. J. Exp. Psychol. 38, 404–411 (1948). 2. Heaton, R. K. Wisconsin Card Sorting Test Manual (Psychological Assesment Resources, Odessa, Florida, 1993). 3. Milner, B. Effects of different brain lesions on card sorting. Arch. Neurol. 9, 90–100 (1963). 4. Drewe, E. A. The effect of type and area of brain lesion on Wisconsin card sorting test performance. Cortex 10, 159–170 (1974). 5. Robinson, A. L., Heaton, R. K., Lehman, R. A. W. & Stilson, D. W. The utility of the Wisconsin Card Sorting Test in detecting and localizing frontal lobe lesions. J. Consult. Clin. Psychol. 48, 605–614 (1980). 6. Janowski, J. S., Shimamura, A. P., Kritchevski, M. & Squire, L. R. Cognitive impairment following frontal damage and its relevance to human amnesia. Behav. Neurosci. 103, 548–560 (1989). 7. Passingham, R. E. Non-reversal shifts after selective prefrontal ablations in

83



article

monkeys (Macaca Mulatta). Neuropsychologia 10, 41–46 (1972). 8. Dias, R., Robbins, T. W. & Roberts, A. C. Dissociation in prefrontal cortex of affective and attentional shifts. Nature 380, 69–72 (1996). 9. Milner, B. in The Frontal Granular Cortex and Behavior (eds Warren, J.M. & Akert, K.) 313–334 (McGraw-Hill, New York, 1964). 10. Mishkin, M. in The Frontal Granular Cortex and Behavior (eds Warren, J.M. & Akert, K.) 219–241 (McGraw-Hill, New York, 1964). 11. Goldman-Rakic, P. S. in Circuitry of primate prefrontal cortex and regulation of behavior by representational memory. Handbook of Physiology, Vol. 2, (eds Mountcastle, V.B., Plum, F. & Geiger, S.R.), 373–417 (American Physiological Society, Bethesda, 1987). 12. Petrides, M. in Handbook of Neuropsychology (eds Boller, F. & Grafman, J.) 75–90 (Elsevier, Amsterdam, 1989). 13. Passingham, R. E. The Frontal Lobes and Voluntary Action (Oxford Univ. Press, Oxford, 1993). 14. Damasio, A. R. On some functions of the human prefrontal cortex. Ann. NY Acad. Sci. 769, 241–251 (1995). 15. Fuster, J. M. The Prefrontal Cortex (Raven, New York, 1997). 16. Rosen, B. R., Buckner, R. L. & Dale, A. M. Event-related functional MRI: Past, present, and the future. Proc. Natl. Acad. Sci. USA 95, 773–780 (1998). 17. Mountain, M. A. & Snow, W. G. Wisconsin Card Sorting Test as a measure of frontal pathology: a review. Clin. Neuropsychol. 7, 108–118 (1993). 18. Friston, K. J., Jezzard, P. & Turner, R. Analysis of functional MRI time-series. Hum. Brain Mapp. 1, 153–171 (1994). 19. Buckner, R. L. et al. Detection of cortical activation during averaged single trials of a cognitive task using functional magnetic resonance imaging. Proc. Natl. Acad. Sci. USA 93, 14878–14883 (1996). 20. Konishi, S. et al. Transient brain activity used in magnetic resonance imaging to detect functional areas. Neuroreport 8, 19–23 (1996). 21. Kim, S. G., Richter, W. & Ugurbil, K. Limitation of temporal resolution in functional MRI. Magn. Reson. Med. 37, 631–636 (1997). 22. Zarahn, E., Aguirre, G. & D’Esposito, M. A trial-based experimental design for fMRI. Neuroimage 6, 122–138 (1997). 23. Blamire, A. M. et al. Dynamic mapping of the human visual cortex by highspeed magnetic resonance imaging. Proc. Natl. Acad. Sci. USA 89, 11069–11073 (1992). 24. Narayan, S. M., Santori, E. M. & Toga, A. W. Mapping functional activity in rodent cortex using optical intrinsic signals. Cereb. Cortex 4, 195–204 (1994). 25. Savoy, R. L. et al. Pushing the temporal resolution of fMRI: studies of very brief visual stimuli, onset variability and asynchrony, and stimulus-correlated changes in noise. Proc SMR 3rd Annual Meeting 450 (1995). 26. Malonek, D. & Grinvald, A. Interactions between electrical activity and cortical microstimulation revealed by imaging spectroscopy: implications for functional brain mapping. Science 272, 551–554 (1996). 27. Buckner, R. L., Raichle, M. E., Miezin, F. M. & Petersen, S. E. Functional anatomic study of memory retrieval for auditory words and visual pictures. J. Neurosci. 16, 6219–6235 (1996). 28. Friston, K. J., Frith, C. D., Liddle, P. F. & Frackowiak, R. S. J. Comparing functional (PET) images: The assessment of significant change. J. Cereb. Blood Flow Metab. 11, 690–699 (1991). 29. Talairach, J. & Tournoux, P. Co-planar stereotaxic atlas of the human brain (Stuttgart, Thieme, 1988).

84

30. Hinke, R. M. et al. Functional magnetic resonance imaging of Broca’s area during internal speech. Neuroreport 4, 675–678 (1993). 31. Kim, K. H. S., Relkin, N. R., Lee, K.-M. & Hirsh, J. Distinct cortical areas associated with native and second languages. Nature 388, 171–174 (1997). 32. Jonides, J. et al. Spatial working memory in humans as revealed by PET. Nature 363, 623–625 (1993). 33. Petrides, M., Alivisatos, B., Meyer, E. & Evans, A. C. Functional activation of the human frontal cortex during the performance of verbal working memory tasks. Proc. Natl. Acad. Sci. USA 90, 878–882 (1993). 34. McCarthy, G. et al. Functional magnetic resonance imaging of human prefrontal cortex activation during a spatial working memory task. Proc. Natl. Acad. Sci. USA 91, 8690–8694 (1994). 35. D’Esposito, M. et al. The neural basis of the central executive system of working memory. Nature 378, 279–281 (1995). 36. Baker, S. C., Frith, C. D., Frackowiak, R. S. J. & Dolan, R. J. Active representation of shape and spatial location in man. Cereb. Cortex 6, 612–619 (1996). 37. Courtney, S. M., Ungerleider, L. G., Keil, K. & Haxby, J. V. Object and spatial visual working memory activate separate neural systems in human cortex. Cereb. Cortex 6, 39–49 (1996). 38. Fiez, J. A. et al. A positron emission tomography study of the short-term maintenance of verbal information. J. Neurosci. 16, 808–822 (1996). 39. Dehaene, S. et al. Cerebral activations during number multiplication and comparison: a PET study. Neuropsychologia 34, 1097–1106 (1996). 40. Cohen, D. C. et al. Temporal dynamics of brain activation during a working memory task. Nature 386, 604–608 (1997). 41. Courtney, S. M., Ungerleider, L. G., Keil, K. & Haxby, J. V. Transient and sustained activity in a distributed neural system for human working memory. Nature 386, 608–611 (1997). 42. Klingberg, T., O’Sullivan, B. T. & Roland, P. E. Bilateral activation of frontoparietal networks by incrementing demand in a working memory task. Cereb. Cortex 7, 465–471 (1997). 43. Prabhakaran, V., Smith, J. A., Desmond, J. E., Glover, G. H. & Gabrieli, J. D. Neural substrates of fluid reasoning: an fMR study of neocortical activation during performance of the Raven’s Progressive Matrices Test. Cog. Psychol. 33, 43–63 (1997). 44. Miyashita, Y. Inferior temporal cortex: Where visual perception meets memory. Ann. Rev. Neurosci. 16, 245–263 (1993). 45. Konishi, S., Nakajima, K., Uchida, I., Sekihara, K. & Miyashita, Y. No-go dominant brain activity in human inferior prefrontal cortex revealed by functional magnetic resonance imaging. Eur. J. Neurosci. 10, 1209–1213 (1998). 46. Berman, K. F. et al. Physiological activation of a cortical network during performance of the Wisconsin Card Sorting Test: a positron emission tomography study. Neuropsychologia 33, 1027–1046 (1995). 47. Nagahama, Y. et al. Cerebral activation during performance of a card sorting test. Brain 119, 1667–1675 (1996). 48. Oldfield, R. C. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9, 97–113 (1971). 49. Sakai, K. et al. Functional mapping of the human somatosensory cortex with echo-planar magnetic resonance imaging. Magn. Reson. Med. 33, 736–743 (1995). 50. Sakai, K. et al. Functional mapping of the human colour centre with echoplanar magnetic resonance imaging. Proc. R. Soc. Lond. B 261, 89–98 (1995).
