We review PET studies of higher-order cognitive processes, including attention (sustained and selective), perception (of objects, faces, and locations), language (word listening, reading, and production), working memory (phonological and visuo-spatial), semantic memory retrieval (intentional and incidental), episodic memory retrieval (verbal and nonverbal), priming, and procedural memory (conditioning and skill learning). For each process, we identify activation patterns including the most consistently involved regions. These regions constitute important components of the network of brain regions that underlie each function.

Primates are unique among mammals in possessing a region of dorsolateral prefrontal cortex with a well-developed internal granular layer. This region is commonly implicated in higher cognitive functions. Despite the histological distinctiveness of primate dorsolateral prefrontal cortex, the work of Rose, Woolsey, and Akert produced a broad consensus among neuroscientists that homologues of primate granular frontal cortex exist in nonprimates and can be recognized by their dense innervation from the mediodorsal thalamic nucleus (MD). Additional characteristics have come to be identified with dorsolateral prefrontal cortex, including rich dopaminergic innervation and involvement in spatial delayed-reaction tasks. However, recent studies reveal that these characteristics are not distinctive of the dorsolateral prefrontal region in primates: MD and dopaminergic projections are widespread in the frontal lobe, and medial and orbital frontal areas may play a role in delay tasks. A reevaluation of rat frontal cortex suggests that the medial frontal cortex, usually considered to be homologous to the dorsolateral prefrontal cortex of primates, actually consists of cortex homologous to primate premotor and anterior cin-date cortex. The lateral MD-projection cortex of rats resembles portions of primate orbital cortex. If prefrontal cortex is construed broadly enough to include orbital and cingulate cortex, rats can be said to have prefrontal cortex. However, they evidently lack homologues of the dorsolateral prefrontal areas of primates. This assessment suggests that rats probably do not provide useful models of human dorsolateral frontal lobe function and dysfunction, although they might prove valuable for understanding other regions of frontal cortex.

The near universally accepted theory that the brain processes information persists in current neural network theory where there is "subsymbolic" computation (Smolensky, 1988) on distributed representations. This theory of brain information processing may suffice for simplifying models simulated in silicon but not for living neural nets where there is ongoing chemical tuning of the input/output transfer function at the nodes, connection weights, network parameters, and connectivity. Here the brain continually changes itself as it intersects with information from the outside. An alternative theory to information processing is developed in which the brain permits and supports "participation" of self and other as constraints on the dynamically evolving, self-organizing whole. The noncomputational process of "differing and deferring" in nonlinear dynamic neural systems is contrasted with Black's (1991) account of molecular information processing. State hyperspace for the noncomputational process of nonlinear dynamical systems, unlike classical systems, has a fractal dimension. The noncomputational model is supported by suggestive evidence for fractal properties of the brain.

The cognition that constructs mental features such as intention, disposition, and character is an aspect of theory of mind. This aspect of representation of minds, which inherently has valence, is viewed from cognitive, evolutionary, and neural perspectives. It is proposed that this cognition is modular, and that it normally operates in association with a valence-free cognition able to represent mental states such as belief. Examples of neural activity capable of supporting the social representations macaque monkeys are believed to possess (understanding of affective displays, purposeful movement, and elemental social interactions) are presented.

During the last two decades there has been a proliferation of studies evaluating the psychophysical and neural attributes of heat-induced pain. Experiments using radiant and contact heat-induced pain have produced different observations thereby broadening our appreciation of the importance of acknowledging how a noxious heat stimulus is delivered; moreover manipulations of stimulus parameters have now provided a foundation for understanding the underlying neural mechanisms of heat-induced pain and their biological significance. The psychophysical attributes of heat-induced pain include highly reliable thresholds for most body regions, minimal adaptation to maintained noxious stimuli, an exquisite sensitivity to small changes in stimulus intensity, slow temporal summation for some types of heat-induced pain (i.e., second pain) but not for others (i.e., first pain), spatial summation---especially for suprathreshold noxious temperatures---and the perceived spread of pain well beyond the actual body area stimulated (i.e., radiation). The present body of information indicates that the pain system is optimally adapted for conveying precise information about intensity, and is less concerned with other stimulus features, such as spatial patterns or boundaries.

It is well established that early in development interconnections within the mammalian visual system are often more widespread and less precise than at maturity. The literature dealing with the formation of visual connections has largely ignored differences in developmental specificity among species differing in their phylogenetic status and/or the visual ecological niche that they occupy. Based on a review of the available evidence, we have formulated an hypothesis to account for the varying degrees of developmental specificity that characterize different visual systems. It is suggested that extremely precise systems required for high-acuity binocular vision exhibit fewer presumed developmental errors than do visual systems characterized by poorer acuity and relatively crude depth perception. The developmental implications of the hypothesis are considered, and specific experiments are proposed to further test its validity.

As part of the general trend toward interdisciplinary research in recent years, a growing number of investigators have come to consider both cognitive and neuroscientific perspectives when theorizing about memory. Although such cognitive neuroscience analyses are a relatively recent development, the approach has precedents in earlier scientific thinking about memory. In this article we present a historical review of three major issues in memory research---consolidation processes, the nature of memory representations, and multiple memory systems. We discuss the nature of the relation between cognitive and neuroscientific approaches to each of these issues with respect to the distinction between collateral, complementary, and convergent relations (Schacter, 1986). Although some early investigators offered analyses that linked psychological and physiological perspectives, there is little historical evidence of systematic or sustained interdisciplinary research. However, more recent work, especially with respect to hypotheses about memory systems, suggests progress toward establishing programmatic interdisciplinary research.

The notion of a neuron that responds selectively to the image of a particular complex object has been controversial ever since Gross and his colleagues reported neurons in the temporal cortex of monkeys that were selective for the sight of a monkey's hand (Gross, Rocha-Miranda, & Bender, 1972). Since that time, evidence has mounted for neurons in the temporal lobe that respond selectively to faces. The present paper presents a critical analysis of the evidence for face neurons and discusses the implications of these neurons for models of object recognition. The paper also presents some possible reasons for the evolution of face neurons and suggests some analogies with the development of language in humans.

After a CS–US learning trial, powerful feedback (FB) selectively modulates the hedonic value of the US, perhaps affecting the CS as well. FB. operates regardless of the subject's attributions or awareness, often influencing the subject's unconscious motives in a single trial. Most of the evidence comes from feeding research where the taste US and the internal FB are independently manipulated, but similar influences are at work with painful, thermal and sexual USs in a wide variety of species including humans. Implications for cognitive theorizing and neural research on mechanisms of learning are discussed.

The literature pertaining to the representation of language in the right hemisphere of hemispherectomy and callosotomy patients is reviewed to ascertain whether it provides an empirical basis for the assertions that (1) the right hemisphere participates in the recovery of language in aphasia, (2) the right hemisphere mediates the reading errors of deep dyslexic and pure alexic patients, and (3) the right hemisphere plays a fixed role in normal reading processes. At present, there appears to be some support for the first assertion, limited support for the second (if individual variation in representation can be accepted), and disconfirming evidence for the third in the data from these populations.

People with profound memory deficits can acquire new motor skills, even though on each day of training they may be unaware of having seen the training apparatus before. Similarly, rabbits can acquire or retain a conditioned nictitating membrane response despite massive lesions of the hippocampus or the cerebral cortex. Both lines of evidence suggest that sub-cortical structures may be sufficient for many forms of motor learning. One possible locus for motor learning is the cerebellum. This article traces the history of our knowledge of structure and function of the cerebellum. The anatomical and physiological evidence demonstrates that the cerebellum has the neural connections necessary to mediate simple forms of motor learning or reflex plasticity. Behavioral studies demonstrate that the cerebellum is involved in modification of the vestibuloocular reflex, recalibration of saccadic eye movements, and acquisition of the conditioned nictitating membrane response. Although the evidence in all three instances suggests that the cerebellum is important, there is no agreement as to whether the cerebellum is always necessary for motor learning or how it might participate. Two views are presented: one supporting the idea of the cerebellum as the locus for motor learning and the other opposing this idea. Some evidence that might resolve these disagreements is discussed.

The neocortex is the most characteristic feature of the human brain. On gross inspection, its convoluted surfaces can be seen to have overgrown and covered most other brain structures. In the functional sphere, it is to the neocortex that we attribute those behaviors assumed to be most uniquely human such as cognition and linguistic behavior. This essay is an attempt to understand how this structure expanded during the course of mammalian evolution. At present, any attempt must be more speculative than definitive, but it is offered in the hope that it will generate more discussion on a topic that is central to all neurobiology, as well as a number of allied disciplines. I will proceed by outlining current views on the evolution of the brain, briefly review the organization of the somatosensory cortex in several mammalian forms, and then discuss in some detail ontogenetic mechanisms that may have some bearing on neocortical phylogeny. The primary proposition put forth is that the mammalian neocortex is relatively unspecified by strict genetic means, and that this allowed the neocortex to expand and adapt to a variety of circumstances during the course of phylogeny.

In the nonmammalian telencephalon, there are neuronal populations corresponding to cell groups in the neocortex of mammals in terms of connections, single unit-responses, chemical content, and functions. Some of these populations in nonmammals, however, are organized in a nonlaminar, rather than laminar fashion. These observations may prompt a reassessment of the functional roles of lamination and the evolutionary origins of the mammalian neocortex. Thus, the role of neural circuits and laminar organization can be differentiated in order to understand the cognitive functions of the neocortex. Moreover, the origins of neocortex can be separable into the precursors of nonlaminar and laminar regions.

Bizarreness is a cognitive feature common to REM sleep dreams, which can be easily measured. Because bizarreness is highly specific to dreaming, we propose that it is most likely brought about by changes in neuronal activity that are specific to REM sleep. At the level of the dream plot, bizarreness can be defined as either discontinuity or incongruity. In addition, the dreamer's thoughts about the plot may be logically deficient. We propose that dream bizarreness is the cognitive concomitant of two kinds of changes in neuronal dynamics during REM sleep. One is the disinhibition of forebrain networks caused by the withdrawal of the modulatory influences of norepinephrine (NE) and serotonin (5HT) in REM sleep, secondary to cessation of firing of locus coeruleus and dorsal raphe neurons. This aminergic demodulation can be mathematically modeled as a shift toward increased error at the outputs from neural networks, and these errors might be represented cognitively as incongruities and/or discontinuities. We also consider the possibility that discontinuities are the cognitive concomitant of sudden bifurcations or “jumps” in the responses of forebrain neuronal networks. These bifurcations are caused by phasic discharge of pontogeniculooccipital (PGO) neurons during REM sleep, providing a source of cholinergic modulation to the forebrain which could evoke unpredictable network responses. When phasic PGO activity stops, the resultant activity in the brain may be wholly unrelated to patterns of activity dominant before such phasic stimulation began. Mathematically such sudden shifts from one pattern of activity to a second, unrelated one is called a bifurcation. We propose that the neuronal bifurcations brought about by PGO activity might be represented cognitively as bizarre discontinuities of dream plot. We regard these proposals as preliminary attempts to model the relationship between dream cognition and REM sleep neurophysiology. This neurophysiological model of dream bizarreness may also prove useful in understanding the contributions of REM sleep to the developmental and experiential plasticity of the cerebral cortex.

Mammals vary in number of visual areas from a few to 20 or more as a result of new visual areas being added to the middle levels of processing hierarchies. Having more visual areas probably increases visual abilities, perhaps in part by allowing more stimulus parameters to be considered. Proposals that each visual area computes and thereby “detects” a specific stimulus attribute have so far dealt with attributes that most mammals can detect and thus do not relate to the issue of species differences in numbers of areas. The problem of forming and maintaining complex patterns of interconnections between many different sets of distinct processing models within an area may limit multiplying functions within a field. In addition, the adding of new visual areas is a way of avoiding constraints on modifying existing visual areas that are imposed by the ongoing functional requirements. Thus, increasing the number of visual or other cortical areas is an effective and apparently common mechanism for evolving new capacities.

The demonstration that functional somatotopic maps within the adult neocortex undergo some degree of reorganization following peripheral injury has aroused considerable interest. The evidence for such reorganization in the rat and monkey is reviewed and it is concluded that in both species there is good evidence for limited functional map reorganization in the adult neocortex following peripheral injury. The significance of such functional map reorganization, particularly in terms of whether or not cortical maps are continuously modifiable throughout life, is discussed. It is concluded that the current evidence for map reorganization is best interpreted in terms of the unmasking of preexisting neuronal circuits rather than as evidence of dynamic cortical selection processes.