Muller, R.U., Stead, M., and Pach, J. (1996). The hippocampus as a cognitive graph. J. Gen. Physiol., 107, 663-694.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... discriminate between different parts of an environment, i.e. they fire only in one part of the environment, but not in others [53] It is generally believed that place cells play an important role in navigation, and a large number of theoretical hippocampus models have been presented (e.g. [54, 73, 46, 6, 4, 48, 65] ) The system of Recce Harris is an implementation of Marr s hippocampus model [42] on a mobile robot. In this model, the hippocampus was viewed as an autoassociative memory which stores a scene representation consisting of the bearings and distances of the surrounding landmarks and of a goal ....

R. U. Muller, M. Stead, and J. Pach. The hippocampus as a cognitive graph. Journal of General Physiology, 107:663 -- 694, 1996.

.... discriminate between different parts of an environment, i.e. they fire only in one part of the environment, but not in others [53] It is generally believed that place cells play an important role in navigation, and a large number of theoretical hippocampus models have been presented (e.g. [4,6,46,48,54,65,73]) The system of Recce and Harris is an implementation of Marr s hippocampus model [42] on a mobile robot. In this model, the hippocampus was viewed as an autoassociative memory which stores a scene representation consisting of the bearings and distances of the surrounding landmarks and of a goal ....

R.U. Muller, M. Stead, J. Pach, The hippocampus as a cognitive graph, Journal of General Physiology 107 (1996) 663--694.

....in the environment. The most direct path in the environment is represented by the set of connected place cells for which the sum of the synaptic weights is the greatest (maximizing the sum of the inverses of the distances) Trullier et al. 48 Progress in Neurobiology As explained by Muller et al. [92], the addition of a few ad hoc mechanisms enable the generation of topological detours and shortcuts. The addition of an obstacle, requiring a detour, is represented by the removal from the graph of the place cells occupied by the obstacle. Thus, paths that previously included these places are no ....

....place node receives the signal, the adjacent place lying on the shortest path corresponds to the adjacent node from which the signal comes from. The direction to this adjacent place is retrieved from the topological graph and the animat moves in this direction. As in the model of Muller et al. [91, 92], the animat is not able to generate metric shortcuts through previously unvisited places, but the spatial representation includes sufficient metric information so that the animat can select between alternate paths on the basis of their lengths. 3.3.3.4 Schmajuk and Thieme (1992) Schmajuk and ....

[Article contains additional citation context not shown here]

R. U. Muller, M. Stead, and J. Pach. The hippocampus as a cognitive graph. Journal of General Physiology, 107:663--694, 1996.

....on hippocampal connectivity described in Section 3.2.2. 5 (i) place cells are sufficiently interconnected so that there are two interconnected place cells for any two given locations in the environment, and (ii) connections have synapses modifiable through a Hebbian rule Muller et al. [91] propose a model of how the hippocampus might implement a topological representation of the environment. Their model also encodes some metric information. The animat exhibits topological navigation and is able to select paths on the basis of the path length, but doesn t exhibit metric navigation ....

....place node receives the signal, the adjacent place lying on the shortest path corresponds to the adjacent node from which the signal comes from. The direction to this adjacent place is retrieved from the topological graph and the animat moves in this direction. As in the model of Muller et al. [91, 92], the animat is not able to generate metric shortcuts through previously unvisited places, but the spatial representation includes sufficient metric information so that the animat can select between alternate paths on the basis of their lengths. 3.3.3.4 Schmajuk and Thieme (1992) Schmajuk and ....

[Article contains additional citation context not shown here]

R. U. Muller, J. L. Kubie, and R. Saypoff. The hippocampus as a cognitive graph (abridged version). Hippocampus, 1(3):243--246, July 1991.

.... (Rolls 1991) However, there is another view, which is increasingly popular, according to which the hippocampus rather functions as an hetero associative network, that learns, stores and recalls the relationships between neighboring places (spatial relationships, e.g. Schmajuk and Thieme, 1992] [Muller et al. 1996]) or between successive events (temporal relationships, e.g. Jensen et al. 1996] Wallenstein and Hasselmo, 1997] At the time of recall, the representation of a place (respectively of an event) associated with the intention of some action, leads to the prediction of the next place ....

.... stemming from dioeerences in the paradigms and the protocols used) It seems however that there is some agreement as to the idea that the hippocampus is involved in dioeerent kinds of memory processes, one of which being spatial memory (O Keefe and Nadel 1978; Nadel 1991) We have followed Muller et al. 1996) in speaking of the hippocampus as a icognitive graphj instead of as a icognitive mapj which is more often used. This allows us not to assume an underlying metric representation which would resemble a imap in the headj (Kuipers 1982) We thus oppose the theory proposed by Touretzky and Redish ....

[Article contains additional citation context not shown here]

Muller, R. U., M. Stead, and J. Pach (1996). The hippocampus as a cognitive graph. Journal of General Physiology 107, 663694.

.... of the rat that are responsive to spatial location (O Keefe and Dostrovsky 1971; O Keefe and Nadel 1978; Wilson and McNaughton, 1993; O Keefe and Burgess, 1996) has led to a number of suggestions about the role of the hippocampus in navigation (O Keefe and Nadel 1978; McNaughton et al. 1991; Muller et al. 1991; Traub et al. 1992; Worden 1992; Hetherington and Shapiro 1993; Burgess et al. 1994; Wan et al. 1994; Blum and Abbott, 1996) A recurring idea is that hippocampal place cells provide an environmental map that aids in navigation (O Keefe and Nadel 1978l; Muller et al. 1991; Traub et al. 1992) It ....

.... McNaughton et al. 1991; Muller et al. 1991; Traub et al. 1992; Worden 1992; Hetherington and Shapiro 1993; Burgess et al. 1994; Wan et al. 1994; Blum and Abbott, 1996) A recurring idea is that hippocampal place cells provide an environmental map that aids in navigation (O Keefe and Nadel 1978l; Muller et al. 1991; Traub et al. 1992) It has been shown that a navigational map can be formed by the potentiation of synapses between hippocampal place cells arising from the exploration of an environment (Blum and Abbott, 1996) This model of formation of a navigational map relies on two key features of the ....

Muller, R. U., Kubie, J. L., and Saypoff, R. (1991) The hippocampus as a cognitive graph.

....a prediction of its future value based on the training experience. The population vector is also shifted for sensory inputs that are not in the training sequence but are similar to training inputs. As we will see from the more detailed computations, this shift is toward the training set (see also Muller et al. 1991). As a result, the network can interpolate to provide predictions for sensory inputs not in the training set. The ability of LTP to produce an ensemble coded prediction on the basis of training experience is seen in figure 1. We consider a two dimensional sensory input vector so that our results ....

Muller, R.U., Kubie, J.L. & Saypoff, R. (1991) The hippocampus as a cognitive graph. Hippocampus 1, 243-246.

.... 1985; Kuipers and Byun, 1991; Mataric, 1992; Nehmzow, 1995) have advocated the use of various types of topological models to represent the connectivity of the environment, and several such models have been devised that aim to mimic known nervous architectures or behavioral capacities in animals (Muller et al. 1991; Mataric, 1992; Schmajuk and Thieme, 1992; Penna and Wu, 1993; Bachelder and Waxman, 1994; Nehmzow, 1995; Scholkopf and Mallot, 1995) Basically, these topological models endow an animat or a robot with cognitive abilities that make possible to recognize the place it is situated in and to ....

....upon the proprioceptive position, orientation, and satisfaction estimates. Such a characteristic is in sharp contrast with many other realizations that implement place recognition capabilities. For instance, several such realizations draw upon biology (Zipser, 1986; Cartwright and Collett, 1987; Muller et al. 1991; Burgess et al. 1994; Bachelder and Waxman, 1994) and implement a neural network in which the firing of some sensory neurons, tuned to the features of some landmarks sensed in a given place, triggers the firing of a specific place cell that codes for this place. Likewise, in (Kuipers and Byun, ....

Muller, R.U., Kubie, J.L. and Saypoff, R. (1991) "The hippocampus as a cognitive graph." Hippocampus 1(3) :243246.

....the place corresponding to the strongest connection weight r. If the prediction signals are too weak, it means either that the goal is far away, or that the spatial representation is incomplete, and in this case, the Action Selection module generates a random movement. The model of Muller et al. [28] also encodes topological links in synapses that connect place cells. Place cells are given a priori and synaptic weights are learned during random exploration. When it moves from one place to another at constant velocity, the corresponding place cells fire with a temporal delay that is ....

....(e.g. 5] or a goal cell (e.g. 4] or a set of coordinates (e.g. 43] or a specific landmark representation (e.g. 32] ffl outputs: how planning is performed, what information is sent to the motor system. When planning is performed, it entails either a classical graph search (e.g. [28]) or a spreading activation process (e.g. 16] The output of most models is a direction of movement, given as a vector (e.g. 12] as a compass direction (e.g. 1] or as a turning angle (e.g. 3] Models that use an a priori place representation usually define the animat s action as that of ....

Muller, R. U., Kubie, J. L., and Saypoff, R. (1991): The hippocampus as a cognitive graph (abridged version). Hippocampus 1(3), 243--246.

.... the fragility of purely metric methods, many researchers resort to the use of various types of topological models that represent the connectivity of the environment, and several such models have been devised that aim to mimic known nervous architectures and behavioral capacities in animals (Muller et al. 1991; Matari c, 1992; Schmajuk and Thieme, 1992; Penna and Wu, 1993; Bachelder and Waxman, 1994; Nehmzow, 1995; Scholkopf and Mallot, 1995) Basically, these topological models endow an animat or a robot with cognitive abilities that make it possible to recognize the place it is situated in and to ....

....upon the proprioceptive position, orientation, and satisfaction estimates. This characteristic is in sharp contrast with many other realizations that implement place recognition capabilities. For instance, several such realizations draw upon biology (Zipser, 1986; Cartwright and Collett, 1987; Muller et al. 1991; Burgess et al. 1994; Bachelder and Waxman, 1994) and implement a neural network in which the firing of some sensory neurons, tuned to the features of some landmarks sensed in a given place, triggers the firing of a specific place cell that codes for this place. Likewise, in Kuipers and Byun ....

Muller, R.U., Kubie, J.L. and Saypoff, R. (1991). The hippocampus as a cognitive graph. Hippocampus, 1(3), 243-246.

....the model immensely. The model also includes local excitation (within reference frame) connections within the hippocampus, as does the multi chart model. However, in the reference frame model this complex connection structure is only assumed to exist after exploration. As has been shown by Muller et al. 1991b, 1996, see also Redish and Touretzky, 1997c, Redish, 1997) this connection structure can be learned by random exploration combined with correlational LTP (i.e. Hebbian learning) Entering a novel environment. When an animal is placed in an environment, we assume that it does not have preconceived ....

R. U. Muller, M. Stead, and J. Pach. The hippocampus as a cognitive graph. Journal of General Physiology, 107(6):663--694, 1996.

....the model immensely. The model also includes local excitation (within reference frame) connections within the hippocampus, as does the multi chart model. However, in the reference frame model this complex connection structure is only assumed to exist after exploration. As has been shown by Muller et al. 1991b, 1996, see also Redish and Touretzky, 1997c, Redish, 1997) this connection structure can be learned by random exploration combined with correlational LTP (i.e. Hebbian learning) Entering a novel environment. When an animal is placed in an environment, we assume that it does not have ....

....inputs from DG. In practice, we have found that most HC cells simulated with this model show at most one place field within a reference frame, but occasionally, some cells do show two subfields. Cells with multiple subfields have been reported in real animals (e.g. O Keefe and Nadel, 1978, Muller et al. 1991a, Wilson and McNaughton, 1994, Markus et al. 1995) As the animal explores the environment, LTP occurs along the learnable (dashed) connections. We assume that this LTP is Hebbian and rectified at 0, so that synaptic strength can only increase. We do not model LTD. Returning to a familiar ....

R. U. Muller, J. L. Kubie, and R. Saypoff. The hippocampus as a cognitive graph. Hippocampus, 1(3):243--246, 1991.

No context found.

Muller, R.U., Stead, M., and Pach, J. (1996). The hippocampus as a cognitive graph. J. Gen. Physiol., 107, 663-694.

No context found.

Muller, R.U., Kubie, J.L., and Saypoff, R. (1991). The hippocampus as a cognitive graph. Hippocampus, 1, 243-246.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC