Olton, David (1943-1994)
OLTON, DAVID (1943-1994)
David Olton was a psychologist who studied the neuroscience of memory and animal cognition. He discovered that rats not only learned and remembered places, but like humans, could keep lists of places in memory for hours. His work on memory in animals, aimed toward modeling human memory and amnesia, led to the discovery that remembering places required the same brain structures in rats and humans: the hippocampal system. These basic discoveries launched a research program into how the hippocampal system normally supported memory, how it was impaired by physical damage, aging, or by diseases such as Alzheimer's dementia, and how impaired memory might be restored.
David Olton was born in New Jersey and reared in Richmond, Virginia. He attended Haverford College, where he enjoyed the questions posed in philosophy classes, but he was more impressed with the surer methods of the natural sciences. After taking a course in psychophysics, he decided that the combination of psychological questions and empirical methods was the right fit. He earned a B.A. in psychology in 1964 and was awarded the Pennsylvania Psychological Association prize for the best undergraduate research project.
Olton's interest in behavioral neuroscience grew with his dissertation work at the University of Michigan, where he studied with Robert L. Isaacson. He wrote his doctoral thesis on the behavioral effects of penicillin injected into the hippocampus. He joined the faculty of Johns Hopkins University at the age of twenty-six in 1969, the same year he was awarded his Ph.D. Within ten years he was a full professor. He served as department chairman from 1982 to 1987 and remained a member of the Johns Hopkins faculty until he was struck down by pancreatic cancer at the age of fifty-one.
Olton's central scientific interest was the brain mechanisms of memory. His career spanned the end of the behaviorist era, the emergence of modern neuroscience and cognitive science, and the widespread acceptance of animal cognition as a legitimate discipline. Some of the key influences on Olton's thinking were the ideas and findings of Scoville and Milner (1957) on case H.M., Tolman (1948) on cognitive maps, Mishkin and Delacour (1975) on primate models of amnesia, Tulving (1972) on episodic memory, Platt (1964) on strong inference, and Garner (1956) on converging operations. A thorough empiricist, Olton made his most important scientific contributions through his rigorous analytic studies of the contribution of the hippocampal system to memory and cognition in rats. He developed novel techniques and pioneered powerful combinations of converging methods. He was a master at translating abstract theoretical concepts into powerful experiments, and he was particularly ingenious when translating human neuropsychological findings into animal models.
The Radial Maze and Working Memory Theory
Olton's most prominent methodological contribution was the radial maze, and his best-known conceptual contribution was the idea that the hippocampus supported working memory. The method and the theory are linked historically and together constitute Olton's most influential work.
Olton and Samuelson introduced the radial maze in 1976. The paper described the maze, an elevated central platform with eight arms extending symmetrically "like spokes from a wheel." In the original task, the end of each arm had a food cup that was filled at the start of a daily trial, and rats learned to forage optimally—to enter each arm once to obtain food without reentering depleted arms. Probe experiments established that rats remembered the spatial location of the arms they had visited to get food and excluded alternative explanations such as scent marking, intramaze cues, and response chaining. This demonstration of spatial memory in rats was a new and important contribution to the emerging field of animal cognition. The result implied that rats represented a list of spatial locations—kept several places in mind—and, through a flexible memory process that was sensitive to interference, remembered which of those locations they visited on a given day. This finding was of such fundamental importance that the radial maze became a symbol of animal cognition.
After establishing these new facts on animal cognition, Olton pursued the neural bases of radial maze performance. Olton first demonstrated that performance depended upon the major extrinsic anatomical connections of the hippocampus (Olton, Walker, and Gage, 1978). Lesions of the entorhinal cortex, the fimbria-fornix, the septum, or the postcommissural fornix produced chance performance on the radial maze. Unilateral lesions of the fornix or entorhinal cortex did not impair performance, whereas crossed lesions that disconnected the hippocampus from cortical and subcortical throughput did. Lesions restricted to hippocampal neurons produced similar impairments to the disconnections (Handelmann and Olton, 1981; Jarrard, 1986). Together, the results were clear: lesions of the hippocampus, its extrinsic connections, or its intrinsic circuitry impaired performance in the radial maze. The task proved to be one of the most sensitive and selective measures of hippocampal function ever devised and is still in use. The basic cognitive requirements of the task—spatial discrimination, recent memory, flexible memory expression, and resistance to interference—influenced Olton's thinking about hippocampal function for his entire career.
Olton's initial analysis of the radial maze emphasized both spatial representation and flexible memory processing. His later work addressed how the hippocampus contributes to each of these two important task demands. He initially emphasized the importance of spatial representation and referred to converging evidence from lesion, stimulation, and recording studies. Thus, he showed that hippocampal seizures impaired radial maze performance (Olton and Wolf, 1982), and that neurons had place fields in the radial maze that, in principle, could represent the different arms as well as a rat's entrance into and exit from those arms (Olton, Branch, and Best, 1978). By the late 1970s, however, Olton became convinced that the hippocampus was crucial for recent memory for a wide range of stimuli aside from spatial ones. The evidence that swayed his view, together with the clarity of Olton's arguments, had a powerful and enduring influence on the neuroscience of memory and hippocampal function.
Olton assigned the arms of a seventeen-arm radial maze into either a "baited" or an "unbaited" set
[Image not available for copyright reasons]
(Olton, 1978; Olton and Papas, 1978). Each arm in the baited set contained food at the start of a trial; the arms in the unbaited set never had food. For the baited set, the contingencies were the same as in the standard radial maze task: during a given trial, the rat had to choose each arm once and to avoid reentering that arm. In contrast, the rat always had to avoid entering the arms in the unbaited set to perform efficiently. The same spatial discrimination ability was required to distinguish arms in both sets, but flexible memory expression was required to remember which of the baited arms had been entered in a given trial. Normal rats entered each baited arm once during each trial and avoided unbaited arms altogether. In contrast, rats trained in the task and then given lesions of the hippocampal system repeatedly reentered baited arms within a trial but avoided entering unbaited arms. Thus, hippocampal lesions impaired flexible memory expression, but not spatial discrimination.
Around 1977, Olton adapted Honig's (1978) terms working memory to emphasize task components that required flexible memory for items within a trial and reference memory to describe task components that were unchanged across trials. Olton and Papas (1978) assembled these findings and concepts to claim that the hippocampus was required for working but not reference memory, even when the items to be remembered were spatial locations. After he investigated a nonspatial working memory task, Olton became more firmly convinced that memory processes better defined the unique contribution of the hippocampal system than spatial representation. Rats were trained to remember which of four visually and tactually distinct arms they had visited in a trial. The location of each of the four arms was changed after each choice, preventing rats from using a spatial representation of their locations. This nonspatial working memory task was severely impaired by lesions of the fimbria-fornix (Olton and Feustle, 1981).
The pattern of intact and impaired performance in the radial maze tasks culminated in Olton's working memory theory of hippocampal function, which informed his most-cited work, Hippocampus, Space, and Memory (Olton, Becker, and Handelmann, 1979). The theory claimed that the hippocampus is required for behaviors that demand working memory, independent of whether the material to be remembered was spatial. Working and reference memory were defined operationally, and testing procedures were explicitly distinguished from memory processes. Olton's working memory theory also addressed one of the two major deficits in amnesia associated with hippocampal damage: the inability of amnesic patients to remember recent events even (and perhaps especially) when those events are comprised of familiar items.
Working Memory: Cognitive Basis and Operational Definitions
Olton intended working-memory tasks to emphasize an event-based, trial-unique memory process (Olton et al., 1979). From the outset, however, the choice of the term working memory posed a problem for students. Cognitive neuroscientists had replaced short-term memory with working memory to describe a memory buffer that served as a representational workspace for manipulating items kept in mind (e.g., Baddeley, 1974). Although he did occasionally describe it as a memory buffer, Olton's working memory was not defined by duration, computational workspace, or consciousness. Rather, Olton used the term to operationalize a memory process described in ethology that was most similar to Tulving's (1972) description of episodic memory (Olton et al., 1979). From Olton's view, working memory entailed memory for events, items that occurred in a specific temporal and individual context. The distinct and varying significance of individual items, encoded as events within a temporal context, provided representations that guide responses more flexibly than their unvarying stimulus content.
Olton's working-memory theory made strong predictions through clear operational definitions. The theory predicted that hippocampal lesions would impair working-memory tasks and spare reference-memory tasks. The clarity and simplicity of the ideas were compelling, and they inspired an international research effort that led to important advances in the neuroscience of memory, as described in detail below. Both of the predictions succeeded often, but not always. For example, some spatial-reference memory tasks did require the hippocampus, most notably in the water maze (Morris et al., 1982). The cumulative data refuted the strongest predictions of working-memory theory, and Olton conceded that this version of the theory was insufficient.
Although the distinction between working and reference memory did not account for the full range of effects of hippocampal system lesions, working memory procedures often required the septohippocampal system and differentially activated hippocampal neurons. In collaboration with Warren Meck, Gary Wenk, and Russ Church, Olton showed that hippocampal lesions impaired working memory for the duration of recently presented stimuli (Meck et al., 1984 ; Olton et al., 1988). In a double dissociation, prefrontal cortical circuits were shown to be crucial for attending to the duration of two simultaneously presented stimuli (Olton, Wenk, Church, and Meck, 1988). Single neurons in the hippocampus were more commonly and more strongly activated during performance of a nonspatial working memory task than during either a spatial or a cued reference-memory task that required the same perceptual discriminations and behaviors (Wible et al., 1986). Memory, rather than other task variables, strongly influenced hippocampal activity. The tasks required discrimination between cue boxes, and lesions of the hippocampal system impaired only the working-memory procedure (Raffaele and Olton, 1988). Seizure stimulation of the hippocampus completely reset working memory (Olton and Wolf, 1981; Knowlton et al., 1985) but had no effect on the learning of a spatial reference-memory discrimination (Knowlton et al., 1989). Spatial working-memory performance in a T-maze provided an especially sensitive, quantitative assay of septo-hippocampal function. Microinjections of GABA agonists or acetylcholine antagonists disrupted working memory performance, reduced acetylcholine release in the hippocampus, and suppressed hippocampal theta in tightly correlated patterns (Givens and Olton, 1990).
As the limitations of the original working-memory theory became clear, Olton sought to correct the principles, logic, and analytic approaches that misdirected the theory. Early in the 1990s he decided that his thinking had been too strongly influenced by categorical interpretations of dissociation experiments (the classic Olton model was a 2x2 table describing a double dissociation), and he began to consider quantitative, parametric experimental designs. His reading of the data had already convinced him that the spatial theory of hippocampal function was incomplete, that working-memory demand was a crucial variable in the extent to which tasks were impaired by hippocampal lesions, and that the categorical, operational definition of working memory limited its usefulness. He also began to reexamine the third cognitive requirement for performing the radial maze task to guide his thinking: resistance to interference (Olton and Shapiro, 1992; Shapiro and Olton, 1994).
Olton's new approach was fruitful. In a nonspatial, continuous, conditional-discrimination task, rats were trained to press one bar if two consecutive stimuli were the same and another bar if they were different (Wan, Pang, and Olton, 1994; see also Wible et al., 1992). By changing the frequency of stimulus repetition and delay interval, the experiment varied both proactive interference and working-memory demand. Even after months of training, high proactive interference and long delays revealed significant non-spatial working-memory impairments in rats with lesions of the hippocampus or fornix (Wan et al., 1994). His new approach to memory research had begun well when he became gravely ill.
Continuing Influence: Theory and Application
Olton's clear predictions and conceptual analyses posed a challenge that persists today: Can any current theory of hippocampal function accurately describe a priori, in operational terms, the full range of tasks that will be impaired or spared by damage to the hippocampal system? The varied-response strategies provided by multiple memory systems suggest that any particular task can be guided by more than one system. Even an operationally defined spatial working memory task in the radial maze can be solved, in principle, by response chaining and thus dispense with the hippocampus. So, operational distinctions aside, what remains of Olton's analysis of hippocampal function?
The conceptual core of working-memory theory emphasized important aspects of hippocampal function. Hindsight makes clear that the original operational definition Olton proposed for working memory did not distinguish between tasks that required memory for recent events (e.g., episodic memory) from those that could be solved by other mechanisms for maintaining short-term memory for recently experienced stimuli or behaviors. The potential complexity of the neural systems underlying operationally defined working memory tasks was known from cognitive neuropsychology, where different working memory systems were dissociated—compare H.M.'s short-term memory for verbal and nonverbal items—(Sidman, Stoddard, and Mohr, 1968), and from research on nonhuman primates (Fuster, 1995). Reference-memory tasks, defined operationally as trialindependent memory, can in principle depend upon memory for repeated episodes as well as those for the rules and procedures that were originally described to be the information-processing core of reference memory tasks (Olton et al., 1979). Subsequent research has revealed some of the complex circuitry underlying working and reference memory.
Research in Norman White's lab at McGill University used the radial maze to show that various operationally defined reference-memory tasks require distinct neural components that support independent memory strategies (e.g., McDonald and White, 1993). The hippocampus, the amygdala, and the caudate nucleus each support a different memory strategy for discriminating arms in the radial maze (Packard, Hirsh, and White, 1989; McDonald and White, 1983). In one such task, the hippocampal system was shown to be necessary for discriminating between adjacent but not separated arms of the maze. To learn this spatial reference memory discrimination, intact rats had to visit the two arms in succession during each daily trial, suggesting that learning the spatial discrimination depended upon comparisons supported by recent working memory (White and Oellet, 1997).
Working-memory mechanisms are also variable in rats, humans, and nonhuman primates. In rats the variety of brain substrates underlying working-memory tasks was shown by Kesner et al. (1993), who reported a triple dissociation among working-memory tasks. Lesions of the caudate, visual cortex, and hippocampus selectively impaired working memory for responses, visual objects, and locations, respectively (Kesner et al., 1993). Olton's latest ideas on hippocampal function, working memory, and interference also continue to influence memory research (e.g. Long and Kesner, 1998; Gilbert et al., 1998; Hampton et al., 1998; Fortin et al., 2002).
Spatial versus Nonspatial Memory
Olton's observations that the hippocampus is not required exclusively for spatial memory have been verified repeatedly. Many nonspatial memory tasks that require flexible responses to events have been shown to require the hippocampus in both rats and nonhuman primates. Lesions of the hippocampus impair social-recognition memory (Kogan et al., 2000), social transmission of food preference (Winocur, 1990; Bunsey and Eichenbaum, 1995), transitive inference (Bunsey and Eichenbaum, 1996; Dusek and Eichenbaum, 1997), DRL performance (Sinden et al., 1986), negative patterning (Sutherland and Rudy, 1989), trace eyelid conditioning (Solomon et al., 1986), and memory for olfactory sequences (Fortin, Agster, and Eichenbaum, 2002). Not all spatial discrimination tasks require the hippocampus. Thus, rats with fornix or hippocampal lesions not only learn to discriminate arms in a radial maze, but they can be trained to find a hidden platform in the Morris water maze using a reference-memory procedure, whereas they are unable to learn flexible responses such as a spatial-reversal learning in the same situation (Eichenbaum, Stewart, and Morris, 1990; Whishaw et al., 1995). Olton argued that by helping to encode the temporal context of events, the hippocampus contributes to behavioral flexibility. Thus, the same stimulus can be approached in one instance and avoided in a second instance because the unique sequence and the outcome of the behavioral interaction with that stimulus are remembered. A widespread view is that the hippocampus is indeed crucial for remembering items in their temporal context and that this memory provides the necessary representation for flexible memory expression. Empirical studies (Fortin et al., 2002) and theoretical reviews (Manns and Squire, 2001; Eichenbaum et al., 1999) emphasize this aspect of hippocampal function, which was the cognitive core of Olton's working memory theory.
Working memory procedures in general, and the radial maze in particular, remain in use in important studies to assess hippocampal function and memory in many species, including mice, nonhuman primates, and humans (Glassman et al., 1994). To cite only a few examples: The tight relationship between place field activity and spatial behavior in rats was first shown in a spatial-working memory task in a four-arm radial maze (O'Keefe and Speakman, 1987). Psychological stress and concomitant elevated corticosteroids were shown to selectively impair working memory in a radial maze (Diamond et al., 1996). Age-related memory deficits in nonhuman primates were examined using a formal variant of the radial-maze task (Rapp et al., 1997). A working-memory procedure in the water maze helped to clarify the importance of hippocampal NMDA receptors in memory (Steele and Morris, 1999). Finally, recent efforts to model Alzheimer's disease in transgenic mice have used a working-memory task in a radial water maze to track the progression of age-related memory deficits (e.g. Arandash et al., 2001). Olton's hypotheses, analyses, and approaches continue to influence memory science.
David Olton's contribution to science included far more than the sum of his research and ideas. Always serious about science, he was informal in his manner and had an ironic sense of humor about himself. These qualities made him a sought collaborator, a valued and respected colleague, and an outstanding mentor. Among his graduate, undergraduate, and postdoctoral students were Fred H. "Rusty" Gage of the Salk Institute, now president of the Society for Neuroscience; John Morrison, director of the Center for Neurobiology at Mount Sinai School of Medicine; and Gary Wenk, professor of psychology at the University of Arizona.
Arendash, G. W., King, D. L., Gordon, M. N., Morgan, D., Hatcher, J. M., Hope, C. E., and Diamond, D. M. (2001). Progressive, age-related behavioral impairments in transgenic mice carrying both mutant amyloid precursor protein and presenilin-1 transgenes. Brain Research 891, 42-53.
Baddeley, A. D., and Hitch, G. (1974). Working memory. In G. A. Bower, ed., Recent advances in learning and motivation. New York: Academic Press.
Bunsey, M., and Eichenbaum, H. (1996). Conservation of hippocampal memory function in rats and humans. Nature 379 255-257.
Diamond, D. M., Fleshner, M., Ingersoll, N., and Rose, G. M. (1996). Psychological stress impairs spatial working memory: Relevance to electrophysiological studies of hippocampal function. Behavioral Neuroscience 110, 661-672.
Eichenbaum, H., Stewart, C., and Morris, R. G. M. (1990). Hippocampal representation in place learning. Journal of Neuroscience 10, 3,531-3,542.
Garner, R. W., Hake, H. W., and Eriksen, C. W. (1956). Operationism and the concept of perception. Psychological Review 63 (3), 149-159.
Givens, B. S., and Olton, D. S. (1990). Cholinergic and GABAergic modulation of medial septal area: Effect on working memory. Behavioral Neuroscience 104, 849-855.
Glassman, R. B., Garvey, K. J., Elkins, K. M., Kasal, K. L., and Couillard, N. L. (1994). Spatial working memory score of humans in a large radial maze, similar to published score of rats, implies capacity close to the magical number 7 ± 2. Brain Research Bulletin 34, 151-159.
Hampton, R. R., Shettleworth, S. J., and Westwood, R. P. (1998). Proactive interference, recency, and associative strength: Comparisons of black-capped chickadees and dark-eyed juncos. Animal Learning and Behavior 26, 475-485.
Handelmann, G. E., and Olton, D. S. (1981). Spatial memory following damage to hippocampal CA3 pyramidal cells with kainic acid: Impairment and recovery with preoperative training. Brain Research 217, 41-58.
Honig, W. K. (1978). Animal memory and animal learning. In H. L. Roitblat, T. G. Bever, and H. S. Terrace, eds., Animal Cognition. Hillsdale, NJ: Erlbaum.
Jarrard, L. E. (1986). Selective hippocampal lesions and behavior: Implications for current research and theorizing. In R. L. Isaacson and K. H. Pribram, eds., The hippocampus, Vol. 4. New York: Plenum.
Kesner, R. P., Bolland, B. L., and Dakis, M. (1993). Memory for spatial locations, motor responses, and objects: Triple dissociation among the hippocampus, caudate nucleus, and extrastriate visual cortex. Experimental Brain Research 93, 462-470.
Kogan, J. H., Frankland, P. W., and Silva, A. J. (2000). Long-term memory underlying hippocampus-dependent social recognition in mice. Hippocampus 10, 47-56.
McDonald, R. J., and White, N. M. (1993). A triple dissociation of memory systems: Hippocampus, amygdala, and dorsal striatum. Behavioral Neuroscience 107, 3-22.
Mishkin, M., and Delacour, J. (1975). An analysis of short-term visual memory in the monkey. Journal of Experimental Psychology [Animal Behavior] 1, 326-334.
Morris, R. G. M., Garrud, P., Rawlins, J. N. P., and O'Keefe, J. (1982). Place navigation impaired in rats with hippocampal lesions. Nature 297, 681-683.
O'Keefe, J., and Speakman, A. (1987). Single unit activity in the rat hippocampus during a spatial memory task. Experimental Brain Research 68, 1-27.
Olton, D. S., Becker, J. T., and Handelmann, G. H. (1979). Hippocampus, space and memory. Behavioral and Brain Sciences 2, 313-365.
Olton, D. S., Branch, M., and Best, P. J. (1978). Spatial correlates of hippocampal unit activity. Experimental Neurology 58, 387-409.
Olton, D. S., and Feustle, W. A. (1981). Hippocampal function re quired for nonspatial working memory. Experimental Brain Research 41, 380-389.
Olton, D. S., and Papas, B. C. (1979). Spatial memory and hippocampal function. Neuropsychologia 17, 669-682.
Olton, D. S., and Samuelson, R. J. (1976). Rememberance of places passed: Spatial memory in rats. Journal of Experimental Psychology: Animal Behavior Processes 2, 97-116.
Olton, D. S., and Shapiro, M. L. (1992). Mnemonic dissociations: The power of parameters. Journal of Cognitive Neuroscience 4, 200-207.
Olton, D. S., Walker, J. A., and Gage, F. H. (1978). Hippocampal connections and spatial discrimination. Brain Research 139, 295-308.
Olton, D. S., Wenk, G. L., Church, R. M., and Meck, W. H. (1988). Attention and the frontal cortex as examined by simultaneous temporal processing. Neuropsychologia 26, 307-318.
Olton, D. S., and Wolf, W. A. (1982). Hippocampal seizures produce retrograde amnesia without a temporal gradient when they reset working memory. Behavioral and Neural Biology 33, 437-452.
Platt, J. R. (1964). Strong inference. Science 146, 347-353.
Raffaele, K. C., and Olton, D. S. (1988). Hippocampal and amygdaloid involvement in working memory for nonspatial stimuli. Behavioral Neuroscience 102, 349-355.
Rapp, P. R., Kansky, M. T., and Roberts, J. A. (1997). Impaired spatial information processing in aged monkeys with preserved recognition memory. NeuroReport 8, 1,923-1,928.
Scoville, W. B., and Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology Neurosurgery and Psychiatry 20, 11-21.
Sidman, M., Stoddard, L. T., and Mohr, J. P. (1968). Some additional quantiative observations of immediate memory in a patient with bilateral hippocampal lesions. Neuropsychologia 6, 245-254.
Steele, R. J., and Morris, R. G. M. (1999). Delay-dependent impairment of a matching-to-place task with chronic and intrahippocampal infusion of the NMDA-antagonist D-AP5. Hippocampus 9, 118-136.
Tolman, E. C. (1948). Cognitive maps in rats and men. Psychological Review 56, 144-155.
Tulving, E. (1972). Episodic and semantic memory. In E. Tulving and W. Donaldson, eds., Organization of memory. New York: Academic Press.
Wan, R. Q., Pang, K., and Olton, D. S. (1994). Hippocampal and amygdaloid involvement in nonspatial and spatial working memory in rats: Effects of delay and interference. Behavioral Neuroscience 108, 866-882.
White, N. M., and Oellet, M.-C. (1997). Roles of movement and temporal factors in spatial learning. Hippocampus 7, 501-510.
Wible, C. G., Findling, R. L., Shapiro, M., Lang, E. J., Crane, S., and Olton, D. S. (1986). Mnemonic correlates of unit activity in the hippocampus. Brain Research 399, 97-110.
Winocur, G. (1990). Anterograde and retrograde amnesia in rats with dorsal hippocampal or dorsomedial thalamic lesions. Behavioural Brain Research 38, 145-154.