Convergence Of Disciplines Propels Cognitive Science

The mind is a difficult thing to understand. And to get a handle on how it works requires not just a biological, but also a psychological and philosophical approach, say those in the field of cognitive science, a 30-year-old discipline that now seems to be making headway in its quest to understand thought and the mind's function. "The bottom line is, we're a bunch of blind men searching for a black cat in a dark room. You grab at anything you can get," says Jonathan D. Cohen, a cognitive neuroscientist at Carnegie Mellon University in Pittsburgh. "We are trying very hard to make the best of the tools we have in order to make headway on a very complex and challenging problem." To understand human consciousness, researchers in formerly disparate fields have to effectively communicate with each other. "The real key is the interaction of disciplines," Cohen says. Cohen works with neural networks, currently a hot topic in cognitive science. Neural networks are mathematical models of how a group of neurons responds to an input. Neural networks are considered successful descriptions of how the brain functions--at least at an elementary level--when they can reproduce the correct output. These neural networks, although conceived of on a mathematical and theoretical level, are translated into computer programs. The assumption underlying neural network studies is that the brain is, in essence, a sophisticated computing device. Traditionally, cognitive scientists thought it possible to understand what a computer program does and how it operates without knowing anything about the machine it runs on, whether it be the brain or a PC, says Cohen. But over the past decade, a growing number of researchers have seen the necessity of understanding the type of machine these brain "programs" run on, he says, and the convergence of those who describe the mind at the behavioral and conceptual level and those who study it at the neuronal level is heating up, Cohen says. As he writes in a 1992 article, "Convergences among these disciplines are beginning to emerge in a number of different areas and in a variety of forms. Neuropsychology, behavioral neurology, computational neuroscience, cognitive neuroscience are all areas in which active integration of the behavioral and neurosciences is now being pursued" (J.D. Cohen, D. Servan- Schreiber, Psychiatric Annals, 22[3]:113-8, 1992). "Much of the recent progress in these areas can be attributed to new methods for simultaneously studying brain function and behavior," Cohen writes. "These include neuroimaging techniques (such as PET and MRI) and electrophysiological techniques (such as event-related potentials) that can be used in humans, as well as more detailed methods that can be used in animal studies (such as single-cell recording and microdialysis). "However, empirical tools alone are not sufficient to define a new area of science. These must be accompanied by a theoretical framework within which findings and ideas can be organized and communicated. In the case of the interface between the cognitive and neurosciences, the framework must be broad enough to be able to capture the details of computations underlying cognition as they are being carried out in the brain." Martha Farah, a cognitive neuropsychologist at the University of Pennsylvania, agrees that biological and psychological experimentalists are discovering an increasing need to find common ground. "There's been a skepticism within cognitive psychology that the kind of descriptions neuroscientists have of the mind's function will ever meet up--that it will be decades or centuries before they meet up. But they are starting to meet up now," she says. A review of scientific literature bears out both Cohen and Farah's assertions that the fields are beginning to develop significant cross-linkages. For instance, just five years ago, several journals in the field, such as Neural Computation and the Journal of Cognitive Neuroscience, had not even appeared. That itself is the mark of a field at once coalescing and rapidly developing. Individual researchers certainly know their own area of expertise: the key people, leading labs, current problems, and general trends. But in a field as variegated as cognitive neuroscience, there are many activities that can escape the attention of even the most wide-ranging researcher. To obtain a global perspective on the broadening shape and scope of cognitive neuroscience, The Scientist drew on the data resources of the Philadelphia-based Institute for Scientific Information (ISI). ISI maintains a database of how often articles in one journal cite articles in another and obtains summary statistics on the frequency or intensity of two-way links between individual journals. The strength of the two-way links is calculated by taking the percentage of all citations of papers in journal A by papers appearing in journal B and the percentage of all citations of journal B's papers by those appearing in journal A, and then calculating the average of the two figures to obtain a measure of re- latedness between journals A and B. Using ISI's program, SCI-MAP, a graphical representation was generated (see accompanying illustration) that represents the interrelationships of journals in the field of cognitive neuroscience. By examining groups of journals that are related to other groups of journals, interconnections between and among fields emerge. In the map, a line between two journals denotes a significant two-way flow of citations; the shorter the line, the stronger the connection. To initiate the map, which has been simplified somewhat here, a few significant journals in the field of cognitive science were chosen, such as Cognition, Cognitive Science, and Behavioral and Brain Sciences. Beginning with one journal, the program then located the journal to which it was most closely linked, and then the next closest linked journal, and so on. Multiple maps were created by cutting off the growth of the map in one direction, which "forced" connections at lower thresholds in other directions. In the end, multiple maps, illustrating various multidisciplinary links, were compiled to obtain this bird's-eye view of the field. Henry Shaffer, a psychologist at the University of Exeter, England, recently called cognitive neuroscience "a meeting place of ideas from computer scientists, psychologists, linguists, philosophers, engineers, neuroscientists, and others interested in the design of intelligent machines" (L.H. Shaffer, Nature, 351:282, 1992). And the accompanying map confirms this perspective. The one missing element from Shaffer's list is philosophy, but only because the database used is restricted to journals in the physical, life, and social sciences. Philosophy journals are covered in ISI's Arts and Humanities Citation Index. The other disciplines are included in ISI's Science Citation Index and Social Sciences Citation Index. The size of the circle representing each publication roughly reflects the number of citations in a given period--in this case, 1989. It, thus, is a depiction of the aggregate influence of a particular journal. In the map, Behavioral and Brain Sciences collects about as many total citations as Cognition, but both are much smaller than Brain Research, Neuroscience, and the Journal of Comparative Neurology. From the journals Cognition, Cognitive Science, and Behavioral and Brain Sciences, one can see several main offshoots: anthropology and sociobiology; artificial intelligence and cybernetics; language, speech, and hearing (and, through these, acoustics); and neurophysiology and neuropsychology. A significant feature of this type of map is that it has no orientation: There is no real significance to a journal or group of journals appearing at the top or bottom, to the right or left. The map is merely a representation of the degrees of relationship among them, and it can be viewed, as it were, from any angle. As cognitive science continues its cross-disciplinary development, it should provide an interesting case study for historians of science. Successive maps of following eras could, in a sense, provide a time-lapse-photography-style snapshot of the field's changing and growing shape. ISI research analyst David Pendlebury provided technical assistance with this article.

Convergence Of Disciplines Propels Cognitive Science

Interested in reading more?

Become a Member of