Knowledge Systems and Material-Specific Memory Deficits
KNOWLEDGE SYSTEMS AND MATERIAL-SPECIFIC MEMORY DEFICITS
The face of cognitive neuroscience has changed drastically since the mid-twentieth century. In the past, lesions were the only basis for inference regarding the functional neuroanatomy of normal cognition. Today the tools of cognitive neuroscience include various methods of neuroimaging, both structural and functional, in normal subjects.
Classifications of Memory Systems
The structure of knowledge representation in the brain is elucidated by the studies of specific dissociations of knowledge loss in brain disease. The question whether there is one memory store or several is of the foremost interest. Studies of amnesias are particularly illuminating in this respect. In most amnesic syndromes, skills are better preserved than facts. Within the declarative domain, context-free information is usually better preserved than context-dependent information. Generic information is better preserved than singular information. Generally, the patient's ability to give a conscious account of previously acquired knowledge is more likely to be impaired than the ability to benefit from this knowledge in various behavioral situations. These observations have been interpreted to indicate the neuropsychological reality of the distinctions between procedural knowledge (skills) and declarative knowledge (facts) (Cohen and Squire, 1980), semantic memory (for general facts) and episodic memory (for personal facts) (Tulving, 1983; Kinsbourne and Wood, 1975), generic knowledge (referring to large classes of equivalent objects) and singular knowledge (referring to unique entities), and explicit knowledge (demonstrated through conscious reports) and implicit knowledge (demonstrated through behavioral gains) (Schacter, 1987; Tulving and Schacter, 1990).
While the phenomenal distinctness of these knowledge categories is widely accepted, consensus is lacking as to whether these knowledge types are mediated by neurally distinct stores, or by different processing demands. Robustness and uniformity of the dissociations are often mentioned as the arguments in favor of separate stores (Schacter, 1985). An alternative hypothesis is that the difference between procedural and declarative, semantic and episodic, generic and singular, and explicit and implicit knowledge types reflects different degrees of accessibility of engrams that are part of the same store. This position suggests that the differences between components of the above dichotomies are quantitative rather than qualitative, and are discrete approximations of continuous variations in the degree of engram accessibility.
Recent studies using PET (positron emission tomography) and fMRI (functional magnetic imaging) methodology have further elucidated the neuroanatomical distinctiveness of these systems. There is evidence to indicate that many aspects of semantic memory processing involve regions of the left lateral prefrontal cortex and the anterior temporal cortex. Some findings converge to suggest that this system is organized hierarchically along the posterior to anterior axis (Martin and Chao, 2001). Functional imaging studies of episodic memory have not provided support for the view that these processes are mediated primarily by the medial temporal lobe memory system (Bookheimer, 1996). Researchers have suggested that encoding and retrieval of episodic information might be related more to the functions of the left and right prefrontal cortex respectively (Tulving et al., 1994), although this view is not consistent with the bulk of information obtained from lesion studies and may be too simplistic to account for the complex nature of these processes.
Modality-Specificity of Knowledge Systems
Two additional types of knowledge-base dissociations have been described: sensory modalities and by semantic categories. Modality-specific knowledge loss is exemplified by associative agnosias and modality-specific aphasias. In associative agnosias, the subject loses the ability to identify objects as members of generic categories (Warrington, 1975; Goldberg, 1990). The deficit may be isolated, in that it may be present without sensory or language impairment, and without dementia. Most important in the context of this analysis, the deficit is modality-specific, and at least three types of associative agnosias have been identified: visual object agnosia (McCarthy and Warrington, 1986), pure astereognosis (Hecaen and Albert, 1978), and auditory associative agnosia (Vignolo, 1982).
In each of these agnosias, the ability to interpret object meaning is impaired with respect to a distinct input modality. A patient can see that a watch is round and flat but does not recognize it as a watch in visual object agnosia; a patient can feel that it has a smooth, glassy surface and a small bump on the side, but does not recognize it as a watch in pure astereognosis; and a patient can hear it tick but does not recognize it as a watch in auditory associative agnosia. The knowledge-base impairment in associative agnosias is evident both in patients' inability to correctly name the object and in their inability to signal its correct meaning through nonverbal means, such as pantomime. However, successful object identification, both verbal and nonverbal, will be accomplished in each of these three conditions with reliance on other sensory modalities. The existence of modality-specific associative agnosias has led to the hypothesis that the nonverbal knowledge base is dimensionalized at least in part by sensory modalities (Warrington, 1975; Goldberg, 1990; Damasio, 1990; Shallice, 1987). This hypothesis is strengthened by the presence of double dissociations between any two types of associative agnosia.
In modality-specific aphasias, the patient can correctly identify the object meaning through nonverbal means (e.g., pantomime) but cannot come up with a correct name (Beauvois, 1982). However, the name is easily retrieved when the patient is allowed to resort to other sensory input modalities. The existence of sense-specific aphasias further supports the notion of modality-specific knowledge stores, by suggesting that each of them has a separate access to an amodal lexical store (Beauvois, 1982).
Modality-specific associative agnosias are distinct not only phenomenally but also neuroanatomically. Each type of agnosia has a distinct cortical territory that is consistent across patients. This observation lends further support to the notion of multiple, neurally distinct, modality-specific knowledge stores. It has been suggested that the modality-specific associative agnosias are all linked predominantly to the left hemisphere (Warrington, 1975; Goldberg, 1990). If this assertion is true, then the left hemisphere emerges as the repository of multiple knowledge systems, verbal and nonverbal alike. It has been further suggested that the neocortical functional organization within the posterior portion of a hemisphere is characterized by continuous, gradiental distributions of cognitive functions. The geometry of the cognitive gradients is determined by the sensory cortices (Goldberg, 1989). This position is consistent with the notion that representations are dimensionalized in terms of sensory modalities. Functional neuroimaging studies in normal subjects have further supported the notion of the distributed nature of mental representations. Information from both PET and fMRI have demonstrated that information about objects and their features may be represented in the same neural systems that are active during their perception (Martin, 2001).
Category-Specific Knowledge Representation
Category-specific knowledge loss has also been reported (Damasio, 1990; Warrington and Shallice, 1990; Hart, Berndt, and Caramazza, 1985). In the lexical domain, this pertains to the double dissociation of comprehension and naming of object names and action names (Goodglass, Klein, Carey, and Jones, 1966; Miceli, Silveri, Villa, and Caramazza, 1984). Further fractionation of noun loss has also been reported (Warrington and Shallice, 1990; Hart, Berndt, and Caramazza, 1985; McKenna and Warrington, 1980).
Category-specific knowledge loss also may manifest itself as a selective inability to describe objects or elicit their mental images (Warrington and Shallice, 1984), or as selective agnosia for certain categories of objects but not for others (Nielsen, 1946). The most common and consistent observation of category-specific knowledge loss is that the knowledge of living objects or foods is more impaired than the knowledge of inanimate objects (Vignolo, 1982; Goldberg, 1989; Hart, Berndt, and Caramazza, 1985; Goodglass et al., 1986). However, researchers have also reported the opposite pattern (Warrington and McCarthy, 1983, 1987).
To account for the overwhelming unidirectionality of dissociation, with most studies reporting greater preservation of knowledge about inanimate than living things, and very few reporting the opposite pattern, it has been proposed that the difference may reflect inherently greater perceptual similarities, and therefore confusability, within the living domain than within the inanimate domain (Riddoch, Humphreys, Coltheart, and Funnell, 1988). Alternatively, it has been proposed that the category-specific knowledge loss may reflect different patterns of relative salience of different sensory modalities for different categories (Goldberg, 1989; Warrington and McCarthy, 1987). The latter helps to explain category-specific double dissociations. It also interrelates category-and sense-specific aspects of mental representations.
The inanimate objects used in most studies are in fact human-made objects or tools. Therefore, it is difficult to know which of the two distinctions, living versus inanimate or human-made versus natural, best captures the observed differences. The latter distinction emphasizes the secondary nature of category-specific aspects relative to modality-specific aspects of knowledge representations. This is because human-made tools have mandatory somatosensory and motor representations in the brain that are absent for most natural objects or foods. Therefore, tools are encoded with reliance on more sensory dimensions compared with most natural objects, which would make the corresponding engrams more robust.
In considering the more esoteric types of category-specific knowledge loss or knowledge preservation (Hart, Berndt, and Caramazza, 1985; Yamadori and Albert, 1973; McKenna and Warrington, 1978), one must also take into account the possible premorbid idiosyncrasies of individual lexical strengths and weaknesses. This may be a potent source of artifact in analyzing postmorbid performance.
Finally, combined category-and modality-specific knowledge loss has been reported in a patient who had a selective loss of living things but not objects in the verbal but not visual domain (McCarthy and Warrington, 1988). Elizabeth Warrington and Tim Shallice (1984) conclude that knowledge is organized along both sensory and category dimensions.
Knowledge of the object's superordinate category is well preserved in modality-specific, category-specific, and combined knowledge loss (Warrington, 1975; McCarthy and Warrington, 1988). This pervasive observation has lent support to the hypothesis that knowledge about things is hierarchic. Researchers have proposed that the access to a specific category member invariably begins with accessing a superordinate category (Warrington, 1975). While this may be true in some cases, the observation of the relative preservation of superordinate knowledge does not in itself necessitate this conclusion. In fact, a different route of object identification has also been proposed, from the basic category to superordinate and subordinate categories (Rosch, 1978).
Researchers have evoked both degraded-store (Warrington and Shallice, 1984) and impaired-access (Humphreys, Riddoch, and Quinlan, 1988) hypotheses to account for category-and modality-specific memory loss. They have suggested that a degraded store is characterized by the uniformity of responses across recall trials, and impaired access by their variability (Warrington and Shallice, 1984; Cermak and O'Connor, 1983; Shallice, 1988). The possible neuroanatomical basis for this distinction may be related to whether the critical lesion affects neocortical sites where representations are distributed, thus resulting in degraded store, or subcortical structures involved in various aspects of activation and arousal, thus resulting in impaired access.
Functional neuroimaging studies in normal subjects also point to the segregation of the neural systems involved in category-specific knowledge. Investigators have shown that specific regions of the ventral temporal cortex respond differentially to processing of various categories (Chao, Haxby, and Martin, 1999). There have been some indications that dissociations in the pattern of activation follow the distinction between animate and inanimate categories (Caramazza and Shelton, 1998); however research findings in the late 1990s and early 2000s have failed to find evidence that these category distinctions exist at the neural level (Devlin et al., 2002).
Beauvois, M. F. (1982). Optic aphasia: A process of interaction between vision and language. Philosophical Transactions of the Royal Society (London) B298, 35-47.
Bookheimer, S. Y. (1996). Functional MRI applications in clinical epilepsy. Neuroimage 4, S139-146.
Caramazza, A., and Shelton, J. R. (1998). Domain-specific knowledge systems in the brain: The animate inanimate distinction. Journal of Cognitive Neuroscience 10, 1-34.
Cermak, L. S., and O'Connor, M. (1983). The retrieval capacity of a patient with amnesia due to encephalitis. Neuropsychologia 21, 213-234.
Chao, L. L., Haxby, J. V., and Martin, A. (1999). Attribute-based neural substrates in temporal cortex for perceiving and knowing about objects. Nature Neuroscience 2, 913-919.
Cohen, N. J., and Squire, L. R. (1980). Preserved learning and retention of pattern-analyzing skill in amnesia: Dissociation of "knowing how" and "knowing that." Science 210, 207-209.
Damasio, A. R. (1990). Category related recognition defects as a clue to the neural substrates of knowledge. Trends in Neurosciences 13, 95-98.
Devlin, J. T., Russell, R. P. Davis, R. H., Price, C. J., Moss, H. E., Fadili, M. J., and Tyler, L. K. (2002). Is there an anatomical basis for category-specificity? Semantic memory studies in PET and fMRI. Neuropsychologia 40, 54-75.
Goldberg, E. (1989). Gradiental approach to the neocortical functional organization. Journal of Clinical and Experimental Neuropsychology 11, 489-517.
—— (1990). Associative agnosias and the functions of the left hemisphere. Journal of Clinical and Experimental Neuropsychology 12, 467-484.
Goodglass, H., Klein, B., Carey, P., and Jones, K. (1966). Specific semantic word categories in aphasia. Cortex 2, 74-89.
Goodglass, H., Wingfield, A., Hyde, M. R., and Theurkauf, J. (1986). Category-specific dissociation in naming and recognition by aphasic patients. Cortex 22, 87-102.
Hart, J., Berndt, R. S., and Caramazza, A. (1985). Category-specific naming deficit following cerebral infarction. Nature 316, 439-440.
Hecaen, H., and Albert, M. L. (1978). Human neuropsychology. New York: Wiley.
Humphreys, G. W., Riddoch, M. J., and Quinlan, P. T. (1988). Cascade processes in picture identification. Cognitive Neuropsychology 5, 67-103.
Kinsbourne, M., and Wood, F. (1975). Short-term memory processes and the amnestic syndrome. In D. Deutsch and J. A. Deutsch, eds., Short-term memory. New York: Academic Press.
Martin, A. (2001). Functional neuroimaging of semantic memory. In R. Cabaza and A. Kingstone, eds., Functional imaging of semantic memory. Cambridge, MA: MIT Press.
McCarthy, R. A., and Warrington, E. K. (1986). Visual associative agnosia: A clinico-anatomical study of a single case. Journal of Neurology, Neurosurgery and Psychiatry 49, 1,233-1,240.
—— (1988). Evidence for modality-specific meaning systems in the brain. Nature 334, 428-430.
McKenna, P., and Warrington, E. K. (1978). Category-specific naming preservation: A single case study. Journal of Neurology, Neurosurgery and Psychiatry 41, 571-574.
—— (1980). Testing for nominal dysphasia. Journal of Neurology, Neurosurgery and Psychiatry 43, 781-788.
Miceli, G., Silveri, M. C., Villa, G., and Caramazza, A. (1984). On the basis for the agrammatic's difficulty in producing main verbs. Cortex 20, 207-220.
Nielsen, J. M. (1946). Agnosia, apraxia, aphasia: Their value in cerebral localization, 2nd edition. New York: Hoeber.
Riddoch, M. J., Humphreys, G. W., Coltheart, M., and Funnell, E. (1988). Semantic systems or system? Neuropsychological evidence re-examined. Cognitive Neuropsychology 5, 3-25.
Rosch, E. (1978). Principles of categorization. In E. Rosch and B. B. Lloyd, eds., Principles of categorization. Hillsdale, NJ: Erlbaum.
Schacter, D. L. (1985). Multiple forms of memory in humans and animals. In N. M. Weinberger, J. L. McGaugh, and G. Lynch, eds., Memory systems of the brain. New York: Guilford Press.
—— (1987). Implicit memory: History and current status. Journal of Experimental Psychology: Learning, Memory, and Cognition 13, 501-518.
Shallice, T. (1987). Impairments of semantic processing: Multiple dissociations. In M. Coltheart, G. Santori, and R. J. Job, eds., The cognitive neuropsychology of language. Hillsdale, NJ: Erlbaum.
—— (1988). From neuropsychology to mental structure. Cambridge, MA: Cambridge University Press.
Tulving, E. (1983). Elements of episodic memory. Oxford: Oxford University Press.
Tulving, E., Kapur, S., Craik, F. I. M., Moscovitch, M., and Houle, S. (1994). Hemispheric encoding/retrieval asymmetry in episodic memory: Positron emission tomography findings. Proceedings of the National Academy of Sciences of the United States of America 91, 2,016-2,020.
Tulving, E., and Schacter, D. L. (1990). Priming and human memory systems. Science 247, 301-306.
Vignolo, L. A. (1982). Auditory agnosia. Philosophical Transactions of the Royal Society (London) B298, 16-33.
Warrington, E. K. (1975). The selective impairment of semantic memory. Quarterly Journal of Experimental Psychology 27, 635-657.
Warrington, E. K., and McCarthy, R. A. (1983). Category specific access dysphasia. Brain 106, 859-878.
—— (1987). Categories of knowledge: Further fractionations and an attempted integration. Brain 110, 1,273-1,296.
Warrington, E. K., and Shallice, T. (1984). Category specific semantic impairments. Brain 107, 829-853.
Yamadori, A., and Albert, M. L. (1973). Word category aphasia. Cortex 9, 112-125.