ELSEVIER PI1 SO361-9230( 96)00146-3 Space and the Hippocampal Formation in Humans ROBIN G. MORRIS,’ ALAN PICKERING, SHARON ABRAHAMS AND JANET D. FEIGENBAUM Brain Research Bulletin, Vol. 40, Nos. 5/6, pp. 487-490, 1996 Copyright 0 1996 Elsevier Science Inc. Printed in the USA. All rights reserved 0361-9230/96 $15.00 + .OO Neuropsychology Unit, Department of Psychology, and Department of Clinical Psychology, institute of Psychiatry, De Crespigny Park, London SE5 8AF, UK [Received 1 October 1995; Accepted 1 February 19961 KEY WORDS: Neurology, Hippocampus, Cognitive maps. INTRODUCTION There is considerable cross species evidence for a specific spatial memory system whose functioning is to maintain the spatial ori- entation of the organism. The contribution of the hippocampal formation to this system has received the main support from find- ings with rhodents, monkeys, and humans [ 2,9]. In rhodents, bilateral lesions of the hippocampus proper result in spatial mem- ory impairment [ 891 and there are many studies demonstrating “ place cells” within the hippocampal formation, which respond specifically when the animal is in a particular location as defined by the spatial configuration of the objects in an environment [ 21. In nonhuman primates, bilateral lesions produce deficits in the delayed response task and electrophysiological recording has identified neurones that respond to particular spatial locations of objects, but also differentially to allocentric and egocentric in- formation, for example, the single-unit recording studies of Fei- genbaum and Rolls [ 51. The above evidence has led to the notion, put forward origi- nally by O’Keefe and Nadel [ 91 that spatial memories are formed using “ cognitive maps,” representions that describe environ- ments, including distances and directions, all of which are stored within the neuronal architecture of the hippocampal formation and related structures. These maps enable a “ viewer indepen- dent” or allocentric representation of the environment to be formed, to facilitate navigation. In humans, the evidence for this is more sparse, but studies have shown selective spatial memory deficits in patients who have undergone neurosurgical lesions of the right temporal lobes for the treatment of intractable epilepsy. This includes studies reporting impairments in recalling the spa- tial locations of objects in an array, and identification of changes in spatial location and spatial composition [11,12]. It seems in humans that the right hippocampal formation specialises in spa- tial memory, in keeping with the more general hemispheric spe- cialisations found in this species. Recently, three studies have been conducted by the current authors using human analogues of animal tasks, and these pro- vide strong support for the role of the hippocampal formation in human spatial memory. These mimic the design of the rat radial arm maze used by Olton and colleagues to test spatial memory in rhodents [lo]. In this task, the rat has to traverse a number of runways that radiate out from a central platform. The essential feature is that when they have reached the end of each runway and received a reward, the rat has to remember not to enter arms that have been previously traversed. Many versions of this maze have been used, showing hippocampal lesioning-induced im- pairments when it is clear that the animal is required to use “ex- tramaze” cues situated around the room to determine their po- sition within the maze. This contrasts with other manipulations, which include the positioning of intramaze cues, situated within each arm of the maze [ 61. Features of this task have been adapted for use with humans, first, using computer presented three-di- mensional graphical spatial displays and, second, by requiring humans subjects to retain knowledge of spatial locations while moving around an arrangement of spatial locations. EXPERIMENTS Experiment I A spatial memory test was designed specifically for this ex- periment, called the EXECUTIVE-GOLF task, developed by Morris and Feigenbaum [4]. The subject is presented with the representation of a golf “ course” on a visual display unit (VDU) Presented in foreground is an array of holes into which the golfer (shown in the distance) must “ putt” a set of balls. The subject can choose a particular hole by touching it, the com- puter recording the position using a touch sensitive screen (Quickbasic software programmed by R. Morris; Intasolve touch sensitive screen interfacing with IBM compatible computer sys- tem). The subject starts by touching the holes in turn until the correct one is found, whereupon the golfer “ putts” the ball into it (a wrong hole is signalled by a warning tone). On subsequent searches, the subject must avoid this location while searching for another correct one. The trial ends when all the holes have been used by golfer and a further trial proceeds (see Fig. 1, which illustrates a set of searches to find eight locations). The task was presented using four, six, or eight holes with four trials at each level. The main outcome measure was the number of times the subject returned to a previously used hole (the between search error). Twenty left and 20 right unilateral temporal lobe resection (TLR) patients were tested on this task. The TLR operation was formed at the Maudsley Hospital, Lon- don, UK, and involves removing 5.5-6.5 cm of mesial temporal tissue from the anterior pole, including the amygdala and anterior ’ Requests for reprints should be addressed to Dr. Robin G. Morris, Head of the Neuropsychology Unit, Institute of Psychiatry, De Crespigny Park, London SE5 8AF, UK 487