- Email: [email protected]
Different neural systems for recognizing plants, animals, and artifacts
Brain Research Bulletin, Vol. 54, No. 3, pp. 313–317, 2001 Copyright © 2001 Elsevier Science Inc. Printed in the USA. All rights reserved 0361-9230/01/$–see front matter
PII S0361-9230(00)00440-8
Different neural systems for recognizing plants, animals, and artifacts Ryuta Kawashima,1,2* Giyoo Hatano,3 Kyoko Oizumi,3 Motoaki Sugiura,1 Hiroshi Fukuda,1,2 Kengo Itoh,4 Takashi Kato,4 Akinori Nakamura,4 Kentaro Hatano4 and Shozo Kojima5 1
IDAC, Tohoku University, Sendai, Japan; 2Aoba Brain Imaging Research Center, Sendai, Japan; 3Graduate School of Social Sciences, Keio University, Tokyo, Japan; 4Department of Biofunctional Research, National Institute for Longevity Sciences, Obu, Japan; and 5Primate Research Institute, Kyoto University, Inuyama, Japan [Accepted 26 October 2000]
ABSTRACT: The purpose of this study was to investigate functional organization in the human brain involved in the representation of knowledge regarding plants. We measured the brain activity of eight male volunteers during the recognition of visual stimuli representing plants, animals and artifacts, using positron emission tomography. The participants were presented with and were required to name silently two different images each of 15 entities belonging to three ontological categories, and 30 series of four to six digits. Marked increases in regional cerebral blood flow were found in the hippocampus and the parahippocampal areas bilaterally and the right lateral occipital cortex during the silent naming of all three categories, compared with that during the silent reading of digits. The right lateral occipital cortex was specifically activated in association with the naming of plants, and the right fusiform cortex was specifically activated in association with the naming of animals. In addition, the right temporo-occipital cortex was activated only during animals and plants, not artifacts. Our results indicate that there were a few characteristic activations for the different categories, and that entities belonging to the different categories are not necessarily represented in different locations of the brain. © 2001 Elsevier Science Inc.
complex visual information in general, and their difficulty is often not limited to the visual modality. However, it is premature to conclude that semantic and conceptual knowledge of living things is represented separately in the brain from that of nonliving things, for two reasons. First, neuroimaging experiments with normal subjects to test the recognition of animals vs. artifacts [5,12,19] have yielded somewhat unclear results, i.e., according to Caramazza and Shelton [3], although all studies show segregation between categories, the specific areas involved differ across studies. These studies have yielded only partially overlapping results, probably because different entities in the animal and tool categories were presented in different forms (photos vs. line drawings) for different tasks (naming vs. samedifferent judgment). Clearly, we are not yet ready to identify the brain areas involved in the processing of information regarding animals and artifacts. Second, even if we did find consistent segregation between the areas for animals and tools, it would still remain debatable as to what categorical contrast is represented. There are at least three plausible ways in which animals vs. artifacts could be categorized, namely, into things that move spontaneously vs. those that do not (i.e., animals vs. non-animals), living vs. non-living, and natural things vs. artifacts. To choose the most suitable among these alternatives, we introduced plants into the experimental design, in addition to animals and artifacts, as members in the category of living things, constituting another ontological class. How is knowledge regarding plants represented in the human brain? This is an interesting issue, but has not been thoroughly studied, except for the cases of vegetables and fruits [7], for the obvious reason that folk knowledge of plants is not very widespread in Western countries, the result is that naming or recognition of plants is a difficult task for ordinary college students. We report in this paper a neuroimaging experiment in which Japanese college students, who supposedly have greater knowledge regarding plants than their Western counterparts, were given the task of naming plants, as well as animals and artifacts. We considered three possible patterns of segregation. First, animals could be represented differently from all other things,
KEY WORDS: PET, Silent naming, Plants, Occipital cortex.
INTRODUCTION Since the publication of two seminal papers by Warrington [25, 26], a growing number of clinical studies have reported that some patients reveal selective impairment of naming or recognizing living things (represented by animals) or non-living things (represented by man-made physical tools). Although some researchers have wondered if such category-specific deficits are spurious and merely reflective of the fact that, while detailed visual processing is needed for recognizing living things, most artifacts can be identified by their functions, recent detailed reports of several cases have made this sensory-functional interpretation of categoryspecific deficits less tenable [3,16,18]. For example, those patients who show much greater difficulty in recognizing animals than artifacts do not necessarily have problems with the processing of
* Address for correspondence: Ryuta Kawashima, M.D., IDAC, Tohoku University, 4-1 Seiryocho, Aobaku, Sendai 980-8575, Japan. Fax: ⫹81-22717-8560; E-mail: [email protected]
313
314
KAWASHIMA ET AL.
Eight right-handed male volunteers participated in the present study. Written informed consent was obtained from each subject on forms approved by the Ethical Committee of the National Institute for Longevity Sciences and the Declaration of Helsinki (1975). All the subjects were healthy, with no past history of psychiatric or neurological illness, and none was on any medication. High-resolution magnetic resonance imaging (MRI) of the brain was performed for each subject.
roughly comparable with those for the animals, plants and artifacts. It was confirmed in other groups consisting of 10 college students, that most of the stimuli could be named without difficulty, and that the average reaction times for the 15 items did not differ significantly among the three categories (1627, 1475, and 1597 ms for the plants, animals and artifacts, respectively; F(2,2) ⫽ 3.21, p ⬎ 0.10). Naming by the participants in the pilot experiment was often at the level of genus species for the plants and animals, and at a comparable level for the artifacts, with some exceptions (e.g., portable TV instead of a television set). The stimuli were presented one by one at an interval of three seconds, at the same location on a head-mounted display (Mediamask; Olympus, Tokyo, Japan). The order of presentation within and between the three categories was randomized across the participants. They were instructed to watch the center of the screen, without shifting their gaze, and to silently say the names of the animals (or plants or tools) in their mind (covert naming). They were encouraged to guess the names, if they were not certain. For the series of digits, they were asked to read them as numbers, for example, “one hundred twenty-three thousand four hundred fiftysix” for 123456. Immediately after the PET experiment, each subject was asked to name aloud each of the entire set of photos.
Positron Emission Tomography
Image Data Analysis
The regional cerebral blood flow (rCBF) was measured using a SIEMENS ECAT EXACT HR positron emission tomography (PET) scanner (Siemens/CTI) in three-dimensional mode, after a bolus injection of H215O (15 mCi per scan, each lasting 120 s). Attenuation-corrected data were reconstructed into 47 image planes with a resulting resolution of 6 mm at the full-width at half maximum. An electric oculogram (EOG) was recorded for each subject during the PET measurements.
In the present study, standard anatomical structures of the human brain atlas (HBA) system of Roland et al. [22] were fitted interactively to each subject’s MRI using both linear and nonlinear parameters. These parameters were subsequently used to transform each subject’s rCBF images and MRI into the standard atlas form. Then, statistical parametric mapping (SPM) software (SPM96; Wellcome Department of Cognitive Neurology, London, UK) was employed to create statistical maps showing any significant changes in relative rCBF [6]. In this study, an isotropic Gaussian filter with a FWHM of 20 mm was used to increase the signal-to-noise ratio and to compensate for individual differences in gyral anatomy. Differences in global flow were covaried using analysis of covariance. Comparisons across conditions were made by means of t-statistics, and thereafter transformed into normally distributed Z-statistics. We used conjunction analysis [20] to determine task-specific activations. The conjunction analysis was applied by masking, whereby the second subtraction was tested only in pixels that reached the significance level (p ⬍ 0.001) in the first subtraction. Masking was performed in two directions, i.e., subtraction of one from the second and vice versa. Because significant rCBF differences were thresholded at p ⬍ 0.001 in both contrasts of these independent conjunctions, the probability of reaching significant conjoint activation by chance was p ⬍ 0.0001 [4]. The conjunction of (plants–animals) with (plants–artifacts), (animals–plants) with (animals–artifacts), and (artifacts–plants) with (artifacts–animals) revealed the specific areas of the brain activated in relation to the naming of plants, animals and artifacts, respectively. The conjunctions of (plants–artifacts) with (animals–artifacts) was also performed to reveal the characteristic activations associated with the naming of living things. The threshold for significance was set at p ⬍ 0.05 (corrected for multiple comparisons) for all the statistical analyses. Finally, each area of activation was superimposed onto the average transformed MRIs of the same eight subjects involved in this study. Anatomical localization of the areas of activation determined in each comparison was made in relation to the mean reformatted MRI.
because humans can easily recognize from an early age, any entity that moves spontaneously [13]. Second, animals and plants could be combined and contrasted with artifacts, because even young children can recognize commonalities between animals and plants [8]. Finally, animals, plants and artifacts could all be represented separately, because they constitute distinct ontological categories, which can be distinguished based upon schemes of the mind [1] or according to specialized mechanisms acquired through evolution [3]. To examine which of the three is the most tenable is not only interesting in itself, but also important for conceptualizing the observed differences in neural correlates between animals and man-made instruments. MATERIALS AND METHODS Subjects
Task Procedures Each subject performed three silent naming tasks (i.e., naming of plants, animals and artifacts), in addition to a control task during the PET measurements. Two different color photographs of each of 15 entities belonging to the three ontological categories were chosen as the visual stimuli for the naming tasks. All the entities were recognized by a single word name in Japanese, known to be a majority of college students. We intentionally chose familiar plants and less familiar animals and artifacts, so that the naming difficulty would not vary greatly among the ontological categories. The plants were bamboo, cactus, cherry, chrysanthemum, dandelion, (field) horsetail, hydrangea, ginkgo, lily, lotus, miscanthus, morning glory, rose, sunflower and tulip. One of the photographs of each plant was a close-up image of a part of the flower or leaf, and the other was an image of the whole plant. The animals were all vertebrates (alligator, camel, kangaroo, chameleon, fox, giraffe, hippopotamus, leopard, ostrich, pelican, penguin, rabbit, rhinoceros, sea otter and squirrel). One of the two photos of each animal was an image of its entire body, and the other, of its face. The artifacts included a balloon, battery, pocket compass, cassette tape, desk lamp, globe, microwave oven, platform scale, radio, rope, signal, stationery case, table, television set and vase. The two photographs of each item differed in size, color, and form where possible. Manually operated tools (e.g., a hammer, a pair of scissors) were excluded, to avoid the activation of areas representing hand movement. As the control task, the students were also asked to silently read 15 series of four to six digits (e.g., 678980, 0687), each presented in two different colors. This range of number of digits was selected to make the reaction times for the numbers
A PET STUDY OF NAMING PLANTS
315 TABLE 1
TALAIRACH COORDINATES AND Z-VALUE OF PEAK ACTIVATION OF ACTIVATED FIELDS THAT SHOWED SIGNIFICANT REGIONAL CEREBRAL BLOOD FLOW (rCBF) INCREASES Z-score Structure
Naming task versus Control Tasks Right temporo-occipital Left lateral occipital Right lateral occipital Right cuneus Right fusiform Right parahippocampal Left hippocampus Left parahippocampal
x
y
z
Plants
Animals
Artifacts
48 ⫺12 14 10 28 28 ⫺16 ⫺38
⫺74 ⫺86 ⫺92 ⫺56 ⫺68 ⫺42 8 ⫺16
14 18 22 12 2 ⫺10 ⫺36 ⫺22
3.9 5.9 5.4 4.4 – 3.9 3.7 4.9
4.7 4.6 – 4.3 – 4.9 4.2 4.4
– 5.4 4.8 4.1 4.1 5.2 4.4 5.0
Stereotaxic coordinates (in mm) identify the location of the maxima of rCBF change corresponding to the atlas of Talairach and Tournoux [24].
RESULTS Behavioral Results The participants seldom failed to name the photos at the postPET inquiry. The naming was not always “correct,” but most incorrect responses consisted of those in which a closely related object was named (e.g., some named the photograph of a leopard as a cheetah or puma) and no incorrect responses crossed the ontological boundaries. Thus, we can conclude that the photographs employed by us evoked the participants’ knowledge of the names of animals, plants and artifacts as intended. The mean (SD) number of saccadic eye movements during animal, plant, artifact and control conditions was 5.5 (4.2), 5.0 (5.8), 5.5 (4.9) and 3.8 (3.7), respectively. The differences were not statistically significant (F ⫽ 0.24, p ⫽ 0.87, analysis of variance). Brain Activation As can be seen in Table 1, marked increases in rCBF were found in the left hippocampus and parahippocampal area, the right parahippocampal area, and the lateral occipital cortex bilaterally in each silent naming condition compared with the silent reading of digits. This was probably because these regions are responsible for controlling the retrieval of names. However, the exact location of the peak activation within these regions differed among the three categories. Although differences among the three naming conditions were smaller than those between the naming conditions and digit reading condition, there were a few characteristic activations for the different categories. The right lateral occipital area was more strongly activated during the naming of plants than during that of animals and artifacts (Fig. 1a). The right parahippocampal area was activated more strongly during the naming of animals than during that of plants and artifacts (Fig. 1b). There was no particular area that was activated more strongly during the naming of artifacts than during that of animals and plants. In addition, the right temporo-occipital area showed higher activation for animals and plants than for artifacts (Fig. 1c). DISCUSSION This experiment was an extension of several previous studies which attempted to segregate the brain activations involved in the
recognition of animals vs. artifacts, to include plants as well. We confirmed that different as well as common areas were activated in association with the naming of plants, animals and artifacts. We found that the naming of plants partially shared activation areas with that of animals or artifacts. Moreover, the results of the experiment strongly suggest that entities belonging to different categories are not necessarily represented in different locations of the brain; rather, they may be represented by different combinations of locations. The exact locations of the activations in this experiment associated with the recognition of animals and artifacts were not exactly the same as those reported by previous neuroimaging studies that assessed picture naming [5,12,19], the results of previous studies were not highly consistent, either. This is probably because a different control condition was used, different members of animal and tool categories were presented, the entities were depicted differently (in color photos), and a different task (silent naming) was used in this study. For example, we intentionally excluded manually operated tools, which may have reduced the activation during the naming of the artifacts in the premotor area and an area in the left middle temporal gyrus. Whether the whole body or its part was presented, and whether or not the name was highly familiar may have affected the locations of activation. An area in the right lateral occipital cortex was specifically activated during the naming of plants in this study. To our knowledge, there have been no studies attempting to show the functional organization of the human brain involved in the representation of knowledge about plants. As discussed in the Introduction, this may be attributable to naming or recognition of plants being too difficult for ordinary college students, because folk knowledge of plants is not very widespread in Western countries [27]. In previous PET studies, activation of adjacent areas was noted in the passive viewing of pictures of familiar objects vs. passive viewing words [15], and naming real objects vs. viewing nonsense objects [12] comparisons indicating that, at least, this specific area in the right lateral occipital cortex is a part of a neural network involved in visually presented object processing. In the present study, an area in the right fusiform cortex was specifically activated during animal naming. In a previous PET study by Damasio et al. [5], brain activations related to animal naming were identified in the left inferior temporal cortex and the left temporal pole. The inconsistency between our results and their results may well be due to the difference in the task employed as
316
KAWASHIMA ET AL.
FIG. 1. (a) Left: A horizontal section of statistical parametric mappings (SPMs) superimposed onto stereotactically normalized T1-weighted magnetic resonance image (MRI) of human brain atlas (HBA) showing plant-specific activation (significantly activated not only in plants vs. artifact, but also in plants vs. animals comparisons). The left sides of the figures correspond to the left hemisphere. (b) Left: A horizontal section of SPMs superimposed onto stereotactically normalized T1-weighted MRI of HBA showing animal specific activation (significantly activated not only in animals vs. artifact, but also in animals vs. plants comparisons). (c) Left: A horizontal section of SPMs superimposed onto stereotactically normalized T1-weighted MRI of HBA showing animal and plant specific activation (significantly activated not only in animals vs. artifact, but also in plants vs. artifacts comparisons). Right: Mean normalized regional cerebral blood flow values during each condition are shown from voxels showing peak activation. Error bars indicate standard deviation.
the control condition. The silent reading of digits was used as the control condition in our study, however, Damasio et al. [5] used a complex control condition using normal and inverted faces accompanied by a judgment task. Our results are in line with the results
of other PET studies which indicate that naming of animals is associated with activation of the fusiform cortex [17], and that the ventral occipito-temporal region was activated more strongly during the recognition of animals than during that of artifacts [10,19].
A PET STUDY OF NAMING PLANTS The other possible explanation of this specific activation is that the right fusiform region was related to the perception of faces, since some photographs of animals showed somewhat human-like faces. It has been argued that presentation of faces is generally associated with activation of the right fusiform gyrus with somewhat less strong activation of the left fusiform gyrus [9,14,21]. The ventral and medial temporal regions of both sides were commonly activated during all the three naming tasks in this study. These activations were, however, more prominent in the left than in the right hemisphere. Consistent results were reported in many other neuroimaging studies employing object identification and picture naming tasks. Sergent et al. [23], Kosslyn et al. [11], and Bookheimer et al. [2], respectively, showed left fusiform and middle temporal activations during the identification of objects vs. gratings, canonical pictures vs. abstract design, and oral naming of pictures vs. abstract designs comparisons. Martin et al. [12] and Kanwisher et al. [9] showed activation of the bilateral fusiform cortex during the naming of real object vs. nonsense objects and familiar or novel vs. scrambled pictures comparisons, respectively. Menard et al. [15] showed activation bilaterally of the inferior and middle temporal cortex during passive viewing of pictures vs. words comparison. Murtha et al. [17] showed bilateral fusiform, bilateral inferior temporal and the right parahippocampal activations during picture naming vs. passive viewing of abstract pattern or of the plus sign. Our results combined with these results of previous studies suggest that the ventral and medial temporal regions are involved in perceptual semantic processing.
317
7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
ACKNOWLEDGEMENTS
This study was supported in part by JSPS-RFTF (97L00202) to RyutaKawashima and Grant-in-Aid for Scientific Research on Priority Areas to Giyoo Hatano (No. 09207105). The authors would like to thank Itaru Tatsumi at the Tokyo Metropolitan Institute for Gerontology and Satoshi Umeda at Keio University for their helpful suggestions.
18. 19. 20.
REFERENCES 1. Atran, S. Folkbiology and the anthropology of science: Cognitive universals and cultural particulars. Behav. Brain Sci. 21:547– 611; 1998. 2. Bookheimer, S.; Zeffiro, T. A.; Blaxton, T.; Gaillard, W.; Theodore, W. Regional cerebral blood flow during object naming and word reading. Hum. Brain Mapp. 3:93–106; 1995. 3. Caramazza, A.; Shelton, J. R. Domain-specific knowledge systems in the brain: The animate-inanimate distinction. J. Cogn. Neurosci. 10: 1–34; 1998. 4. Cornette, L.; Dupont, P.; Spileers, W.; Sunaert, S.; Michiels, J.; Van Hecke, P.; Mortelmans, L.; Orban, G. A. Human cerebral activity evoked by motion reversal and motion onset. Brain 121:143–157; 1998. 5. Damasio, H.; Grabowski, T. J.; Tranel, D.; Hichwa, R. D.; Damasio, A. R. A neural basis for lexical retrieval. Nature 380:499 –505; 1996. 6. Friston, K. J.; Holmes, A. P.; Worsley, K. J.; Poline, J.-P.; Frith, C. D.;
21.
22.
23. 24. 25. 26. 27.
Frackowiak, R. S. J. Statistical parametric maps in functional imaging: A general linear approach. Hum. Brain Mapp. 2:189 –210; 1995. Hillis, A. E.; Caramazza, A. Category-specific naming and comprehension impairment: A double dissociation. Brain 114:2081–2094; 1991. Inagaki, K.; Hatano, G. Young children’s recognition of commonalties between animals and plants. Child Dev. 67:2823–2840; 1996. Kanwisher, N.; McDermott, J.; Chun, M. M. The fusiform face area: A module in human extrastriate cortex specialized for face perception. J. Neurosci. 17:4302– 4311; 1997. Kohler, S.; Kapur, S.; Moscovitch, M.; Winocur, G.; Houle, S. Dissociation of pathways for object and spatial vision: A PET study in humans. Neuroreport 6:1865–1868; 1995. Kosslyn, S. M.; Alpert, N. M.; Thompson, W. L.; Chabris, C. E.; Rauch, S. L.; Anderson, A. K. Identifying objects seen from different viewpoints: A PET investigation. Brain 117:1055–1071; 1994. Martin, A.; Wiggs, C. L.; Ungerleider, L. G.; Haxby, J. V. Neural correlates of category-specific knowledge. Nature 379:649 – 652; 1996. Massey, C. M.; Gelman, R. Preschooler’s ability to decide whether a photographed unfamiliar object can move itself. Dev. Psychol. 24: 307–317; 1998. McCarthy, G.; Puce, A.; Gore, J. C.; Allison, T. Face-specific processing in the human fusiform gyrus. J. Cogn. Neurosci. 9:605– 610; 1997. Menard, M. T.; Kosslyn, S. M.; Thompson, W. L.; Alpert, N. M.; Rauch, S. L. Encoding words and pictures: A positron emission tomography study. Neuropsychology 34:185–194; 1996. Moss, H. E.; Tyler, L. K.; Jenings, F. When leopards lose the irspots: Knowledge of visual properties in category-specific deficits for living things. Cogn. Neuropsychol. 14:901–950; 1997. Murtha, S.; Chertkow, H.; Beauregard, M.; Evans, A. The neural substrate of picture naming. J. Cogn. Neurosci. 11:399 – 423; 1999. Norwood, N. J.; Ablex, H. J. J.; Berndt, R. S.; Caramazza, A. Category-specific naming deficit following cerebral infarction. Nature 316: 439 – 440; 1985. Perani, D.; Cappa, S. F.; Bettinardi, V.; Bressi, S.; Gorno-Tempini, M.; Matarrese, M.; Fazio, F. Different neural systems for the recognition of animals and man-made tools. Neuroreport 6:1637–1641; 1995. Price, C. J.; Friston, K. J. Cognitive conjunction: A new approach to brain activation experiments. Neuroimage 5:261–270; 1996. Puce, A.; Allison, T.; Asgari, M.; Gore, J. C.; McCarthy, G. Differential sensitivity of human visual cortex to faces, letterstrings, and textures: A functional magnetic resonance imaging study. J. Neurosci. 16:5205–5215; 1996. Roland, P. E.; Graufelds, C. J.; Wahlin, J.; Ingelman, L.; Andersson, M.; Ledberg, A.; Pedersen, J.; Akerman, S.; Dabringhaus, A.; Zilles, K. Human brain atlas: For high resolution functional and anatomical mapping. Hum. Brain Mapp. 1:173–184; 1994. Sergent, J.; Ohta, S.; MacDonald, B. Functional neuroanatomy of face and object processing. Brain 115:15–36; 1992. Talairach, J.; Tournoux, P. Co-planar stereotaxic atlas of the human brain. Stuttgart: George Thieme Verlag; 1998. Warrington, E. K.; McCarthy, R. Category specific access dysphasia. Brain 106:859 – 878; 1983. Warrington, E. K.; Shallice, T. Category specific semantic impairments. Brain 107:829 – 854; 1984. Wolff, P.; Medin, D. L.; Pankratz, C. Evolution and devolution of folkbiolgical knowledge. Cognition 73:177–204; 1999.
PII S0361-9230(00)00440-8
Different neural systems for recognizing plants, animals, and artifacts Ryuta Kawashima,1,2* Giyoo Hatano,3 Kyoko Oizumi,3 Motoaki Sugiura,1 Hiroshi Fukuda,1,2 Kengo Itoh,4 Takashi Kato,4 Akinori Nakamura,4 Kentaro Hatano4 and Shozo Kojima5 1
IDAC, Tohoku University, Sendai, Japan; 2Aoba Brain Imaging Research Center, Sendai, Japan; 3Graduate School of Social Sciences, Keio University, Tokyo, Japan; 4Department of Biofunctional Research, National Institute for Longevity Sciences, Obu, Japan; and 5Primate Research Institute, Kyoto University, Inuyama, Japan [Accepted 26 October 2000]
ABSTRACT: The purpose of this study was to investigate functional organization in the human brain involved in the representation of knowledge regarding plants. We measured the brain activity of eight male volunteers during the recognition of visual stimuli representing plants, animals and artifacts, using positron emission tomography. The participants were presented with and were required to name silently two different images each of 15 entities belonging to three ontological categories, and 30 series of four to six digits. Marked increases in regional cerebral blood flow were found in the hippocampus and the parahippocampal areas bilaterally and the right lateral occipital cortex during the silent naming of all three categories, compared with that during the silent reading of digits. The right lateral occipital cortex was specifically activated in association with the naming of plants, and the right fusiform cortex was specifically activated in association with the naming of animals. In addition, the right temporo-occipital cortex was activated only during animals and plants, not artifacts. Our results indicate that there were a few characteristic activations for the different categories, and that entities belonging to the different categories are not necessarily represented in different locations of the brain. © 2001 Elsevier Science Inc.
complex visual information in general, and their difficulty is often not limited to the visual modality. However, it is premature to conclude that semantic and conceptual knowledge of living things is represented separately in the brain from that of nonliving things, for two reasons. First, neuroimaging experiments with normal subjects to test the recognition of animals vs. artifacts [5,12,19] have yielded somewhat unclear results, i.e., according to Caramazza and Shelton [3], although all studies show segregation between categories, the specific areas involved differ across studies. These studies have yielded only partially overlapping results, probably because different entities in the animal and tool categories were presented in different forms (photos vs. line drawings) for different tasks (naming vs. samedifferent judgment). Clearly, we are not yet ready to identify the brain areas involved in the processing of information regarding animals and artifacts. Second, even if we did find consistent segregation between the areas for animals and tools, it would still remain debatable as to what categorical contrast is represented. There are at least three plausible ways in which animals vs. artifacts could be categorized, namely, into things that move spontaneously vs. those that do not (i.e., animals vs. non-animals), living vs. non-living, and natural things vs. artifacts. To choose the most suitable among these alternatives, we introduced plants into the experimental design, in addition to animals and artifacts, as members in the category of living things, constituting another ontological class. How is knowledge regarding plants represented in the human brain? This is an interesting issue, but has not been thoroughly studied, except for the cases of vegetables and fruits [7], for the obvious reason that folk knowledge of plants is not very widespread in Western countries, the result is that naming or recognition of plants is a difficult task for ordinary college students. We report in this paper a neuroimaging experiment in which Japanese college students, who supposedly have greater knowledge regarding plants than their Western counterparts, were given the task of naming plants, as well as animals and artifacts. We considered three possible patterns of segregation. First, animals could be represented differently from all other things,
KEY WORDS: PET, Silent naming, Plants, Occipital cortex.
INTRODUCTION Since the publication of two seminal papers by Warrington [25, 26], a growing number of clinical studies have reported that some patients reveal selective impairment of naming or recognizing living things (represented by animals) or non-living things (represented by man-made physical tools). Although some researchers have wondered if such category-specific deficits are spurious and merely reflective of the fact that, while detailed visual processing is needed for recognizing living things, most artifacts can be identified by their functions, recent detailed reports of several cases have made this sensory-functional interpretation of categoryspecific deficits less tenable [3,16,18]. For example, those patients who show much greater difficulty in recognizing animals than artifacts do not necessarily have problems with the processing of
* Address for correspondence: Ryuta Kawashima, M.D., IDAC, Tohoku University, 4-1 Seiryocho, Aobaku, Sendai 980-8575, Japan. Fax: ⫹81-22717-8560; E-mail: [email protected]
313
314
KAWASHIMA ET AL.
Eight right-handed male volunteers participated in the present study. Written informed consent was obtained from each subject on forms approved by the Ethical Committee of the National Institute for Longevity Sciences and the Declaration of Helsinki (1975). All the subjects were healthy, with no past history of psychiatric or neurological illness, and none was on any medication. High-resolution magnetic resonance imaging (MRI) of the brain was performed for each subject.
roughly comparable with those for the animals, plants and artifacts. It was confirmed in other groups consisting of 10 college students, that most of the stimuli could be named without difficulty, and that the average reaction times for the 15 items did not differ significantly among the three categories (1627, 1475, and 1597 ms for the plants, animals and artifacts, respectively; F(2,2) ⫽ 3.21, p ⬎ 0.10). Naming by the participants in the pilot experiment was often at the level of genus species for the plants and animals, and at a comparable level for the artifacts, with some exceptions (e.g., portable TV instead of a television set). The stimuli were presented one by one at an interval of three seconds, at the same location on a head-mounted display (Mediamask; Olympus, Tokyo, Japan). The order of presentation within and between the three categories was randomized across the participants. They were instructed to watch the center of the screen, without shifting their gaze, and to silently say the names of the animals (or plants or tools) in their mind (covert naming). They were encouraged to guess the names, if they were not certain. For the series of digits, they were asked to read them as numbers, for example, “one hundred twenty-three thousand four hundred fiftysix” for 123456. Immediately after the PET experiment, each subject was asked to name aloud each of the entire set of photos.
Positron Emission Tomography
Image Data Analysis
The regional cerebral blood flow (rCBF) was measured using a SIEMENS ECAT EXACT HR positron emission tomography (PET) scanner (Siemens/CTI) in three-dimensional mode, after a bolus injection of H215O (15 mCi per scan, each lasting 120 s). Attenuation-corrected data were reconstructed into 47 image planes with a resulting resolution of 6 mm at the full-width at half maximum. An electric oculogram (EOG) was recorded for each subject during the PET measurements.
In the present study, standard anatomical structures of the human brain atlas (HBA) system of Roland et al. [22] were fitted interactively to each subject’s MRI using both linear and nonlinear parameters. These parameters were subsequently used to transform each subject’s rCBF images and MRI into the standard atlas form. Then, statistical parametric mapping (SPM) software (SPM96; Wellcome Department of Cognitive Neurology, London, UK) was employed to create statistical maps showing any significant changes in relative rCBF [6]. In this study, an isotropic Gaussian filter with a FWHM of 20 mm was used to increase the signal-to-noise ratio and to compensate for individual differences in gyral anatomy. Differences in global flow were covaried using analysis of covariance. Comparisons across conditions were made by means of t-statistics, and thereafter transformed into normally distributed Z-statistics. We used conjunction analysis [20] to determine task-specific activations. The conjunction analysis was applied by masking, whereby the second subtraction was tested only in pixels that reached the significance level (p ⬍ 0.001) in the first subtraction. Masking was performed in two directions, i.e., subtraction of one from the second and vice versa. Because significant rCBF differences were thresholded at p ⬍ 0.001 in both contrasts of these independent conjunctions, the probability of reaching significant conjoint activation by chance was p ⬍ 0.0001 [4]. The conjunction of (plants–animals) with (plants–artifacts), (animals–plants) with (animals–artifacts), and (artifacts–plants) with (artifacts–animals) revealed the specific areas of the brain activated in relation to the naming of plants, animals and artifacts, respectively. The conjunctions of (plants–artifacts) with (animals–artifacts) was also performed to reveal the characteristic activations associated with the naming of living things. The threshold for significance was set at p ⬍ 0.05 (corrected for multiple comparisons) for all the statistical analyses. Finally, each area of activation was superimposed onto the average transformed MRIs of the same eight subjects involved in this study. Anatomical localization of the areas of activation determined in each comparison was made in relation to the mean reformatted MRI.
because humans can easily recognize from an early age, any entity that moves spontaneously [13]. Second, animals and plants could be combined and contrasted with artifacts, because even young children can recognize commonalities between animals and plants [8]. Finally, animals, plants and artifacts could all be represented separately, because they constitute distinct ontological categories, which can be distinguished based upon schemes of the mind [1] or according to specialized mechanisms acquired through evolution [3]. To examine which of the three is the most tenable is not only interesting in itself, but also important for conceptualizing the observed differences in neural correlates between animals and man-made instruments. MATERIALS AND METHODS Subjects
Task Procedures Each subject performed three silent naming tasks (i.e., naming of plants, animals and artifacts), in addition to a control task during the PET measurements. Two different color photographs of each of 15 entities belonging to the three ontological categories were chosen as the visual stimuli for the naming tasks. All the entities were recognized by a single word name in Japanese, known to be a majority of college students. We intentionally chose familiar plants and less familiar animals and artifacts, so that the naming difficulty would not vary greatly among the ontological categories. The plants were bamboo, cactus, cherry, chrysanthemum, dandelion, (field) horsetail, hydrangea, ginkgo, lily, lotus, miscanthus, morning glory, rose, sunflower and tulip. One of the photographs of each plant was a close-up image of a part of the flower or leaf, and the other was an image of the whole plant. The animals were all vertebrates (alligator, camel, kangaroo, chameleon, fox, giraffe, hippopotamus, leopard, ostrich, pelican, penguin, rabbit, rhinoceros, sea otter and squirrel). One of the two photos of each animal was an image of its entire body, and the other, of its face. The artifacts included a balloon, battery, pocket compass, cassette tape, desk lamp, globe, microwave oven, platform scale, radio, rope, signal, stationery case, table, television set and vase. The two photographs of each item differed in size, color, and form where possible. Manually operated tools (e.g., a hammer, a pair of scissors) were excluded, to avoid the activation of areas representing hand movement. As the control task, the students were also asked to silently read 15 series of four to six digits (e.g., 678980, 0687), each presented in two different colors. This range of number of digits was selected to make the reaction times for the numbers
A PET STUDY OF NAMING PLANTS
315 TABLE 1
TALAIRACH COORDINATES AND Z-VALUE OF PEAK ACTIVATION OF ACTIVATED FIELDS THAT SHOWED SIGNIFICANT REGIONAL CEREBRAL BLOOD FLOW (rCBF) INCREASES Z-score Structure
Naming task versus Control Tasks Right temporo-occipital Left lateral occipital Right lateral occipital Right cuneus Right fusiform Right parahippocampal Left hippocampus Left parahippocampal
x
y
z
Plants
Animals
Artifacts
48 ⫺12 14 10 28 28 ⫺16 ⫺38
⫺74 ⫺86 ⫺92 ⫺56 ⫺68 ⫺42 8 ⫺16
14 18 22 12 2 ⫺10 ⫺36 ⫺22
3.9 5.9 5.4 4.4 – 3.9 3.7 4.9
4.7 4.6 – 4.3 – 4.9 4.2 4.4
– 5.4 4.8 4.1 4.1 5.2 4.4 5.0
Stereotaxic coordinates (in mm) identify the location of the maxima of rCBF change corresponding to the atlas of Talairach and Tournoux [24].
RESULTS Behavioral Results The participants seldom failed to name the photos at the postPET inquiry. The naming was not always “correct,” but most incorrect responses consisted of those in which a closely related object was named (e.g., some named the photograph of a leopard as a cheetah or puma) and no incorrect responses crossed the ontological boundaries. Thus, we can conclude that the photographs employed by us evoked the participants’ knowledge of the names of animals, plants and artifacts as intended. The mean (SD) number of saccadic eye movements during animal, plant, artifact and control conditions was 5.5 (4.2), 5.0 (5.8), 5.5 (4.9) and 3.8 (3.7), respectively. The differences were not statistically significant (F ⫽ 0.24, p ⫽ 0.87, analysis of variance). Brain Activation As can be seen in Table 1, marked increases in rCBF were found in the left hippocampus and parahippocampal area, the right parahippocampal area, and the lateral occipital cortex bilaterally in each silent naming condition compared with the silent reading of digits. This was probably because these regions are responsible for controlling the retrieval of names. However, the exact location of the peak activation within these regions differed among the three categories. Although differences among the three naming conditions were smaller than those between the naming conditions and digit reading condition, there were a few characteristic activations for the different categories. The right lateral occipital area was more strongly activated during the naming of plants than during that of animals and artifacts (Fig. 1a). The right parahippocampal area was activated more strongly during the naming of animals than during that of plants and artifacts (Fig. 1b). There was no particular area that was activated more strongly during the naming of artifacts than during that of animals and plants. In addition, the right temporo-occipital area showed higher activation for animals and plants than for artifacts (Fig. 1c). DISCUSSION This experiment was an extension of several previous studies which attempted to segregate the brain activations involved in the
recognition of animals vs. artifacts, to include plants as well. We confirmed that different as well as common areas were activated in association with the naming of plants, animals and artifacts. We found that the naming of plants partially shared activation areas with that of animals or artifacts. Moreover, the results of the experiment strongly suggest that entities belonging to different categories are not necessarily represented in different locations of the brain; rather, they may be represented by different combinations of locations. The exact locations of the activations in this experiment associated with the recognition of animals and artifacts were not exactly the same as those reported by previous neuroimaging studies that assessed picture naming [5,12,19], the results of previous studies were not highly consistent, either. This is probably because a different control condition was used, different members of animal and tool categories were presented, the entities were depicted differently (in color photos), and a different task (silent naming) was used in this study. For example, we intentionally excluded manually operated tools, which may have reduced the activation during the naming of the artifacts in the premotor area and an area in the left middle temporal gyrus. Whether the whole body or its part was presented, and whether or not the name was highly familiar may have affected the locations of activation. An area in the right lateral occipital cortex was specifically activated during the naming of plants in this study. To our knowledge, there have been no studies attempting to show the functional organization of the human brain involved in the representation of knowledge about plants. As discussed in the Introduction, this may be attributable to naming or recognition of plants being too difficult for ordinary college students, because folk knowledge of plants is not very widespread in Western countries [27]. In previous PET studies, activation of adjacent areas was noted in the passive viewing of pictures of familiar objects vs. passive viewing words [15], and naming real objects vs. viewing nonsense objects [12] comparisons indicating that, at least, this specific area in the right lateral occipital cortex is a part of a neural network involved in visually presented object processing. In the present study, an area in the right fusiform cortex was specifically activated during animal naming. In a previous PET study by Damasio et al. [5], brain activations related to animal naming were identified in the left inferior temporal cortex and the left temporal pole. The inconsistency between our results and their results may well be due to the difference in the task employed as
316
KAWASHIMA ET AL.
FIG. 1. (a) Left: A horizontal section of statistical parametric mappings (SPMs) superimposed onto stereotactically normalized T1-weighted magnetic resonance image (MRI) of human brain atlas (HBA) showing plant-specific activation (significantly activated not only in plants vs. artifact, but also in plants vs. animals comparisons). The left sides of the figures correspond to the left hemisphere. (b) Left: A horizontal section of SPMs superimposed onto stereotactically normalized T1-weighted MRI of HBA showing animal specific activation (significantly activated not only in animals vs. artifact, but also in animals vs. plants comparisons). (c) Left: A horizontal section of SPMs superimposed onto stereotactically normalized T1-weighted MRI of HBA showing animal and plant specific activation (significantly activated not only in animals vs. artifact, but also in plants vs. artifacts comparisons). Right: Mean normalized regional cerebral blood flow values during each condition are shown from voxels showing peak activation. Error bars indicate standard deviation.
the control condition. The silent reading of digits was used as the control condition in our study, however, Damasio et al. [5] used a complex control condition using normal and inverted faces accompanied by a judgment task. Our results are in line with the results
of other PET studies which indicate that naming of animals is associated with activation of the fusiform cortex [17], and that the ventral occipito-temporal region was activated more strongly during the recognition of animals than during that of artifacts [10,19].
A PET STUDY OF NAMING PLANTS The other possible explanation of this specific activation is that the right fusiform region was related to the perception of faces, since some photographs of animals showed somewhat human-like faces. It has been argued that presentation of faces is generally associated with activation of the right fusiform gyrus with somewhat less strong activation of the left fusiform gyrus [9,14,21]. The ventral and medial temporal regions of both sides were commonly activated during all the three naming tasks in this study. These activations were, however, more prominent in the left than in the right hemisphere. Consistent results were reported in many other neuroimaging studies employing object identification and picture naming tasks. Sergent et al. [23], Kosslyn et al. [11], and Bookheimer et al. [2], respectively, showed left fusiform and middle temporal activations during the identification of objects vs. gratings, canonical pictures vs. abstract design, and oral naming of pictures vs. abstract designs comparisons. Martin et al. [12] and Kanwisher et al. [9] showed activation of the bilateral fusiform cortex during the naming of real object vs. nonsense objects and familiar or novel vs. scrambled pictures comparisons, respectively. Menard et al. [15] showed activation bilaterally of the inferior and middle temporal cortex during passive viewing of pictures vs. words comparison. Murtha et al. [17] showed bilateral fusiform, bilateral inferior temporal and the right parahippocampal activations during picture naming vs. passive viewing of abstract pattern or of the plus sign. Our results combined with these results of previous studies suggest that the ventral and medial temporal regions are involved in perceptual semantic processing.
317
7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
ACKNOWLEDGEMENTS
This study was supported in part by JSPS-RFTF (97L00202) to RyutaKawashima and Grant-in-Aid for Scientific Research on Priority Areas to Giyoo Hatano (No. 09207105). The authors would like to thank Itaru Tatsumi at the Tokyo Metropolitan Institute for Gerontology and Satoshi Umeda at Keio University for their helpful suggestions.
18. 19. 20.
REFERENCES 1. Atran, S. Folkbiology and the anthropology of science: Cognitive universals and cultural particulars. Behav. Brain Sci. 21:547– 611; 1998. 2. Bookheimer, S.; Zeffiro, T. A.; Blaxton, T.; Gaillard, W.; Theodore, W. Regional cerebral blood flow during object naming and word reading. Hum. Brain Mapp. 3:93–106; 1995. 3. Caramazza, A.; Shelton, J. R. Domain-specific knowledge systems in the brain: The animate-inanimate distinction. J. Cogn. Neurosci. 10: 1–34; 1998. 4. Cornette, L.; Dupont, P.; Spileers, W.; Sunaert, S.; Michiels, J.; Van Hecke, P.; Mortelmans, L.; Orban, G. A. Human cerebral activity evoked by motion reversal and motion onset. Brain 121:143–157; 1998. 5. Damasio, H.; Grabowski, T. J.; Tranel, D.; Hichwa, R. D.; Damasio, A. R. A neural basis for lexical retrieval. Nature 380:499 –505; 1996. 6. Friston, K. J.; Holmes, A. P.; Worsley, K. J.; Poline, J.-P.; Frith, C. D.;
21.
22.
23. 24. 25. 26. 27.
Frackowiak, R. S. J. Statistical parametric maps in functional imaging: A general linear approach. Hum. Brain Mapp. 2:189 –210; 1995. Hillis, A. E.; Caramazza, A. Category-specific naming and comprehension impairment: A double dissociation. Brain 114:2081–2094; 1991. Inagaki, K.; Hatano, G. Young children’s recognition of commonalties between animals and plants. Child Dev. 67:2823–2840; 1996. Kanwisher, N.; McDermott, J.; Chun, M. M. The fusiform face area: A module in human extrastriate cortex specialized for face perception. J. Neurosci. 17:4302– 4311; 1997. Kohler, S.; Kapur, S.; Moscovitch, M.; Winocur, G.; Houle, S. Dissociation of pathways for object and spatial vision: A PET study in humans. Neuroreport 6:1865–1868; 1995. Kosslyn, S. M.; Alpert, N. M.; Thompson, W. L.; Chabris, C. E.; Rauch, S. L.; Anderson, A. K. Identifying objects seen from different viewpoints: A PET investigation. Brain 117:1055–1071; 1994. Martin, A.; Wiggs, C. L.; Ungerleider, L. G.; Haxby, J. V. Neural correlates of category-specific knowledge. Nature 379:649 – 652; 1996. Massey, C. M.; Gelman, R. Preschooler’s ability to decide whether a photographed unfamiliar object can move itself. Dev. Psychol. 24: 307–317; 1998. McCarthy, G.; Puce, A.; Gore, J. C.; Allison, T. Face-specific processing in the human fusiform gyrus. J. Cogn. Neurosci. 9:605– 610; 1997. Menard, M. T.; Kosslyn, S. M.; Thompson, W. L.; Alpert, N. M.; Rauch, S. L. Encoding words and pictures: A positron emission tomography study. Neuropsychology 34:185–194; 1996. Moss, H. E.; Tyler, L. K.; Jenings, F. When leopards lose the irspots: Knowledge of visual properties in category-specific deficits for living things. Cogn. Neuropsychol. 14:901–950; 1997. Murtha, S.; Chertkow, H.; Beauregard, M.; Evans, A. The neural substrate of picture naming. J. Cogn. Neurosci. 11:399 – 423; 1999. Norwood, N. J.; Ablex, H. J. J.; Berndt, R. S.; Caramazza, A. Category-specific naming deficit following cerebral infarction. Nature 316: 439 – 440; 1985. Perani, D.; Cappa, S. F.; Bettinardi, V.; Bressi, S.; Gorno-Tempini, M.; Matarrese, M.; Fazio, F. Different neural systems for the recognition of animals and man-made tools. Neuroreport 6:1637–1641; 1995. Price, C. J.; Friston, K. J. Cognitive conjunction: A new approach to brain activation experiments. Neuroimage 5:261–270; 1996. Puce, A.; Allison, T.; Asgari, M.; Gore, J. C.; McCarthy, G. Differential sensitivity of human visual cortex to faces, letterstrings, and textures: A functional magnetic resonance imaging study. J. Neurosci. 16:5205–5215; 1996. Roland, P. E.; Graufelds, C. J.; Wahlin, J.; Ingelman, L.; Andersson, M.; Ledberg, A.; Pedersen, J.; Akerman, S.; Dabringhaus, A.; Zilles, K. Human brain atlas: For high resolution functional and anatomical mapping. Hum. Brain Mapp. 1:173–184; 1994. Sergent, J.; Ohta, S.; MacDonald, B. Functional neuroanatomy of face and object processing. Brain 115:15–36; 1992. Talairach, J.; Tournoux, P. Co-planar stereotaxic atlas of the human brain. Stuttgart: George Thieme Verlag; 1998. Warrington, E. K.; McCarthy, R. Category specific access dysphasia. Brain 106:859 – 878; 1983. Warrington, E. K.; Shallice, T. Category specific semantic impairments. Brain 107:829 – 854; 1984. Wolff, P.; Medin, D. L.; Pankratz, C. Evolution and devolution of folkbiolgical knowledge. Cognition 73:177–204; 1999.