Place cell references

Transcription

1 Alexandrov, Y.I., Grinchenko, Y.V., Laukka, S., Jarvilehto, T., Maz, V.N. & Korpusova, A.V. (1993). Effect of ethanol on hippocampal neurons depends on their behavioural specialization. Acta Physiologica Scandinavica 149: Brown, E.N., Frank, L.M., Tang, D., Quirk, M.C. & Wilson, M.A. (1998). A statistical paradigm for neural spike train decoding applied to position prediction from ensemble firing patterns of rat hippocampal place cell. Journal of Neuroscience 18(18): The problem of predicting the position of a freely foraging rat based on the ensemble firing patterns of place cells recorded from the CA1 region of its hippocampus is used to develop a two-stage statistical paradigm for neural spike train decoding. In the first, or encoding stage, place cell spiking activity is modeled as an inhomogeneous Poisson process whose instantaneous rate is a function of the animal's position in space and phase of its theta rhythm. The animal's path is modeled as a Gaussian random walk. In the second, or decoding stage, a Bayesian statistical paradigm is used to derive a nonlinear recursive causal filter algorithm for predicting the position of the animal from the place cell ensemble firing patterns. The algebra of the decoding algorithm defines an explicit map of the discrete spike trains into the position prediction. The confidence regions for the position predictions quantify spike train information in terms of the most probable locations of the animal given the ensemble firing pattern. Under our inhomogeneous Poisson model position was a three to five times stronger modulator of the place cell spiking activity than theta phase in an open circular environment. For animal 1 (2) the median decoding error based on 34 (33) place cells recorded during 10 min of foraging was 8.0 (7.7) cm. Our statistical paradigm provides a reliable approach for quantifying the spatial information in the ensemble place cell firing patterns and defines a generally applicable framework for studying information encoding in neural systems. Anderson, M.I. & Jeffery, K.J. (2001). Interaction of sensory cues in the control of place cell remapping. Society for Neuroscience Abstracts 31(744.5): 383. Aota, Y., Yamaguchi, Y., Lipa, P., Sofukiu, T. & McNaughton, B.L. (2001). The two components of theta phase precession in rat hippocampal neurons. Society for Neuroscience Abstracts 31(643.18): 330. Hippocampal place cells in freely running rats exhibit a characteristic relation between the position and the firing phase with respect to the theta rhythm. Yamaguchi and McNaughton (1998) proposed that the distribution consists of two components, one with strong correlation Page 1

2 between phase and position (C1) and the other with no correlation (C2). In the present paper, quantitative analyses, based on the two components hypothesis, of the distribution of spikes in the phase-position plane were carried out for several experimental data sets in which the animals performed different simple appetitively motivated spatial tasks on a linear track, a circular track and triangular track. The probability density of spikes in the phase-position plane was quantified by a function consisting of two normal functions by means of the EM algorithm. The results show that the two components hypothesis is well applicable in each data set. The C1 at DG is advanced to CA1 in phase by 0.2 theta cycle (as shown in the figure), which strongly suggests that C1 is generated at an earlier stage than DG and projected to CA region. C2 is found only in CA1 and its magnitude differs in different cells. Furthermore, the correlation between the two components suggests some functional relevance of their interactions. Supported by: JST CREST & MH01565 Barbieri, R., Frank, L.M., Quirk, M.C., Wilson, M.A. & Brown, E.N. (2000). Construction and analysis of the inhomogeneous general inverse Gaussian probability model of place cell spiking activity. Society for Neuroscience Abstracts 30. Neural systems represent information about stimuli from the outside world in the stochastic structure of their firing patterns. Accurate characterization of this stochastic structure is crucial for deciphering how neural systems encode and transmit information. We introduce a new description of hippocampal place cell spiking activity as a function of position based on the inhomogeneous general inverse Gaussian (IGIG) probability model. This model class has the inhomogeneous Poisson (IP), gamma (IG) and inverse Gaussian (IIG) probability densities as special cases. We apply this model to hippocampal place cells recorded from rats running in an open field environment and to both hippocampal and entorhinal place cells recorded from rats running in a linear environment. Q-Q and K-S plot goodness-of-fit methods based on the time-rescaling theorem provide a readily interpretable, graphical summaries of the model fits to the spike train data. Both the IG and IIG give significantly better descriptions of place cell spiking activity than the IP model; the former in its representation of long ISI s and the latter in its representation of bursts. The IGIG model offers a further significant improvement by combining these characteristics of both the IG and IIG models. These findings thus far suggest a flexible framework for modeling place cell spiking activity under varied experimental conditions. Supported by: US NIMH grants MH59733 and MH61637 Barbieri, R., Frank, L.M., Quirk, M.C., Wilson, M.A. & Brown, E.N. (2001). Page 2

3 Measuring the precision of spatial information representation by decoding ensemble place cell spiking activity. Society for Neuroscience Abstracts 31(953.3). Position prediction based on the ensemble firing patterns of CA1 place cells may yield important insights on how neural populations encode information. Our previous decoding algorithm assumed the animals path to be a random walk and that place cell spiking activity obeyed an inhomogeneous Poisson (IP) model with rate as a function of position. Goodness-of-fit analysis showed that path increments are more consistent with a low-order autoregressive (AR) model than a random walk, and that an inhomogeneous gamma (IG) model gave a more accurate description of place cell spiking activity than IP. We present a new decoding algorithm to predict the animal's path from the spike train ensemble based on a nonlinear recursive causal filter constructed using the AR and IG models. The simultaneous activity of 37 place cells was recorded from a Long-Evans rat foraging in an open, circular environment for 23 minutes, and the animals position was measured at 30 Hz. We estimated the AR and IG model parameters from the first 13 minutes and then applied the decoding algorithms to the last 10 minutes of data. Our decoding analysis illustrates how the neural representation of position can be decoded from the ensemble firing patterns, and shows that better predictions (accuracy and coverage probability) of the animals position in the open environment can be made from the ensemble firing pattern of as few as 30 place cells by modeling the non- Poisson spiking activity behavior and the correlation structure in the animals path. Supported by: Grants NIMH MH59733, MH61637, and NSF IBN Best, P.J., White, A.M. & Minai, A. (2001). Spatial processing in the brain: The activity of hippocampal place cells. Annual Reviews of Neuroscience 24: Bezzi, M., Leutgeb, S., Treves, A. & Mizumori, S.J.Y. (2000). Information analysis of location-selective cells in hippocampus and lateral septum. Society for Neuroscience Abstracts 30. In addition to location-specific representations of the environment, hippocampus receives and encodes in its principal cells information about movement. Lateral septal cells also code for location and movement. Information analysis was used to investigate the coding properties of simultaneously recorded hippocampal and lateral septal cells. In both areas, synergy and redundancy was observed for the encoding of location and velocity. Information for each property was differentially affected in the absence of visual cues. For hippocampal principal cells, the information for both velocity and position was decreased in darkness; velocity coding was affected to a larger extent compared to Page 3

4 position coding. In contrast, lateral septal cells only showed a decrease in information for position while velocity information remained unaffected by the changes in the visual environment. This dissociation in the encoding of the two types of information may be a consequence of convergent hippocampal projections to lateral septum or due to other afferent information to lateral septum. In addition, we randomized the temporal structure of spike trains in hippocampus and lateral septum while leaving the spatial structure of the rate coding for position intact. This analysis confirmed that differences in the distribution of the crosscorrelation coefficients between light and dark are not directly related to the changes in the spatial properties of septal and hippocampal cells, but related to the temporal relation of single-unit activity in both areas. The relative timing of neural activity may therefore be critical for the selective transfer of hippocampal rate codes for location and velocity to subcortical structures. Supported by: HFSP RG B and MH58755 Bland, B.H. & Oddie, S.D. (2001). Theta band oscillation and synchrony in the hippocampal formation and associated structures: The case for its role in sensorimotor integration. Review. Behavioral Brain Research 127(1-2): The current review advances the argument that it is naive to ascribe a unitary function to the hippocampal formation (HPC). Rather, it is more productive to consider the hippocampal formation as consisting of a number of subsystems, each subsystem defined by its own particular neural circuitry. Among examples of neural circuitry appearing in current hippocampal literature are theta, beta and gamma oscillations, sharp waves, place cells and head orientation cells. Data are reviewed supporting the case that theta band oscillation and synchrony is involved in mechanisms underlying sensorimotor integration. Specifically, the neural circuitry underlying the production of oscillation and synchrony (theta) in limbic cortex and associated structures function in the capacity of providing voluntary motor systems with continually updated feedback on their performance relative to changing environmental (sensory) conditions. A crucial aspect of this performance is the intensity with which the motor programs are initiated and maintained. The ascending brainstem HPC synchronizing pathways make the primary contribution in this regard. These pathways originate in the rostral pontine region, ascend and synapse with caudal diencephalic nuclei, which in turn send projections to the medial septal region. The medial septum functions as the node in the ascending pathways, sending both cholinergic and GABA-ergic projections to the HPC. An updated version of the sensorimotor integration model including anatomical details is presented and discussed. Page 4

5 Booth, V. & Bose, A. (2001). Neural mechanisms for generating rate and temporal codes in model CA3 pyramidal cells. Journal of Neurophysiology 85(6): The effect of synaptic inhibition on burst firing of a two-compartment model of a CA3 pyramidal cell is considered. We show that, depending on its timing, a short dose of fast decaying synaptic inhibition can either delay or advance the timing of firing of subsequent bursts. Moreover, increasing the strength of the inhibitory input is shown to modulate the burst profile from a full complex burst, to a burst with multiple spikes, to single spikes. We additionally show how slowly decaying inhibitory input can be used to synchronize a network of pyramidal cells. Implications for the phase precession phenomenon of hippocampal place cells and for the generation of temporal and rate codes are discussed. Bose, A. & Recce, M. (2001). Phase precession and phase-locking of hippocampal pyramidal cells. Hippocampus 11(3): We propose that the activity patterns of CA3 hippocampal pyramidal cells in freely running rats can be described as a temporal phenomenon, where the timing of bursts is modulated by the animal's running speed. With this hypothesis, we explain why pyramidal cells fire in specific spatial locations, and how place cells phase-precess with respect to the EEG theta rhythm for rats running on linear tracks. We are also able to explain why wheel cells phase-lock with respect to the theta rhythm for rats running in a wheel. Using biophysically minimal models of neurons, we show how the same network of neurons displays these activity patterns. The different rhythms are the result of inhibition being used in different ways by the system. The inhibition is produced by anatomically and physiologically diverse types of interneurons, whose role in controlling the firing patterns of hippocampal cells we analyze. Each firing pattern is characterized by a different set of functional relationships between network elements. Our analysis suggests a way to understand these functional relationships and transitions between them. Bower, M.R., Euston, D.R., Gebara, N.M. & McNaughton, B.L. (2001). The role of the hippocampus in disambiguating context in a sequence task. Society for Neuroscience Abstracts 31(316.7). Theories of how sequences are encoded in the brain have postulated an asymmetric, associative strengthening of connections between cells representing sequential elements. One difficulty with this notion concerns the problem of disambiguating the sequential context of repeated segments. For example, to retrieve the sequential relationship ABCDCEF, it would be necessary to encode the element C differently in its two sequential contexts Page 5

6 (i.e., to "orthogonalize" the two representations of C). Wood et al. (2000) convincingly showed this effect in the hippocampus of rats running a simple alternation task on a "T" maze. Many pyramidal cells fired selectively in the stem of the "T" depending on the direction of the upcoming turn. These data strongly suggest that the hippocampus per se may provide the essential disambiguating code. Experiments addressing the same problem were ongoing in this lab at the time of the Wood et al. report. Rats were trained to run for rewarding brain stimulation to a sequence of goal locations marked by illuminated diodes, located around the perimeter of a 1.5 meter diameter circular arena (e.g., the rat ran to numbered zones ). After training, rats were required to run the sequence without cues. Rats acquired novel sequences within minute sessions and showed few errors even on the repeated segments dorsal CA1 pyramidal cells were recorded during each experimental session. In contrast to the results of Wood et al., place fields were indistinguishable on all repeated segments. It appears that orthogonalization of repeated elements within the hippocampus may not be necessary for disambiguating the sequential context of repeated segments. Supported by: NS20331, MH01565 & JST CREST Brown, E.N., Nguyen, D.P., Frank, L.M., Wilson, M.A. & Solo, V. (2001). An analysis of neural receptive field plasticity by point process adaptive filtering. Proceedings of the National Academy of Sciences USA 98(21): Neural receptive fields are plastic: with experience, neurons in many brain regions change their spiking responses to relevant stimuli. Analysis of receptive field plasticity from experimental measurements is crucial for understanding how neural systems adapt their representations of relevant biological information. Current analysis methods using histogram estimates of spike rate functions in nonoverlapping temporal windows do not track the evolution of receptive field plasticity on a fine time scale. Adaptive signal processing is an established engineering paradigm for estimating time-varying system parameters from experimental measurements. We present an adaptive filter algorithm for tracking neural receptive field plasticity based on point process models of spike train activity. We derive an instantaneous steepest descent algorithm by using as the criterion function the instantaneous log likelihood of a point process spike train model. We apply the point process adaptive filter algorithm in a study of spatial (place) receptive field properties of simulated and actual spike train data from rat CA1 hippocampal neurons. A stability analysis of the algorithm is sketched in the. The adaptive algorithm can update the place field parameter estimates on a millisecond time scale. It reliably tracked the migration, changes in scale, and Page 6

7 changes in maximum firing rate characteristic of hippocampal place fields in a rat running on a linear track. Point process adaptive filtering offers an analytic method for studying the dynamics of neural receptive fields. Brun, V.H., Otnaess, M.K., Witter, M.P., Moser, M.B. & Moser, E.I. (2001). Place representation in hippocampal area CA1 in the absence of input from area CA3. Society for Neuroscience Abstracts 31(643.6): 329. The indirect pathway to CA1 from the entorhinal cortex (EC) through the trisynaptic circuit has been regarded as the main excitatory input to CA1. However, the CA1 also receives a strong direct projection from layer III neurons of the EC. We examined whether the direct input is sufficient for establishing and maintaining location-specific activity in CA1 pyramidal cells. To disconnect CA3 from CA1, localized knife-cuts were made along the septo-temporal axis of the dorsal hippocampus, and tetrodes were then implanted in the dorsal CA1. The contralateral hippocampus was removed by ibotenic acid. CA1 pyramidal cell activity was recorded while the rats were walking on a linear track or in a square black box (1 m2) with a white cue card on one of the sidewalls. CA1 cells showed place fields both on the treadmill and in the box, and many fields remained stable between trials and days. Histological examination revealed that the cut had separated the CA1 from the CA3 along most of the dorsal hippocampus. Fluorogold injections at the recording site showed no or only a few retrogradely labeled neurons in the CA3. These results suggest that the indirect pathway to CA1 through CA3 is not necessary for establishing and maintaining place fields, although it cannot be excluded yet that the remaining very small subset of CA3 axons is sufficient for establishing and maintaining place fields in single CA1 pyramidal cells. Supported by: EU (QLG3-CT ), Norw. Res. Council and T. Erbo's Foundation Burgess, N. & O'Keefe, J. (1996). Cognitive graphs, resistive grids, and the hippocampal representation of space. Review. Journal of General Physiology 107(6): Burgess, N. & O'Keefe, J. (1996). Neuronal computations underlying the firing of place cells and their role in navigation. Hippocampus 6(6): Our model of the spatial and temporal aspects of place cell firing and their role in rat navigation is reviewed. The model provides a candidate mechanism, at the level of individual cells, by which place cell information concerning self-localization could be used to guide navigation to previously visited reward sites. The model embodies specific predictions regarding the formation of place fields, the phase coding of place cell Page 7

8 firing with respect to the hippocampal theta rhythm, and the formation of neuronal population vectors downstream from the place cells that code for the directions of goals during navigation. Recent experiments regarding the spatial distribution of place cell firing have confirmed our initial modeling hypothesis, that place fields are formed from Gaussian tuning curve inputs coding for the distances from environmental features, and enabled us to further specify the functional form of these inputs. Other recent experiments regarding the temporal distribution of place cell firing in two-dimensional environments have confirmed our predictions based on the temporal aspects of place cell firing on linear tracks. Directions for further experiments and refinements to the model are outlined for the future. Burgess, N., Becker, S., King, J.A. & O'Keefe, J. (2001). Memory for events and their spatial context: Models and experiments. Review. Philosophical Transactions of the Royal Society (London), Series B: Biological Sciences 356(1413): The computational role of the hippocampus in memory has been characterized as: (i) an index to disparate neocortical storage sites; (ii) a timelimited store supporting neocortical long-term memory; and (iii) a contentaddressable associative memory. These ideas are reviewed and related to several general aspects of episodic memory, including the differences between episodic, recognition and semantic memory, and whether hippocampal lesions differentially affect recent or remote memories. Some outstanding questions remain, such as: what characterizes episodic retrieval as opposed to other forms of read-out from memory; what triggers the storage of an event memory; and what are the neural mechanisms involved? To address these questions a neural-level model of the medial temporal and parietal roles in retrieval of the spatial context of an event is presented. This model combines the idea that retrieval of the rich context of reallife events is a central characteristic of episodic memory, and the idea that medial temporal allocentric representations are used in long-term storage while parietal egocentric representations are used to imagine, manipulate and re-experience the products of retrieval. The model is consistent with the known neural representation of spatial information in the brain, and provides an explanation for the involvement of Papez's circuit in both the representation of heading direction and in the recollection of episodic information. Two experiments relating to the model are briefly described. A functional neuroimaging study of memory for the spatial context of life-like events in virtual reality provides support for the model's functional localization. A neuropsychological experiment suggests that the hippocampus does store an allocentric representation of spatial locations. Page 8

9 Burgess, N., Donnett, J.G. & O'Keefe, J. (1998). The representation of space and the hippocampus in rats, robots and humans. Z Naturforsch 53(7-8): Experimental evidence suggests that the hippocampus represents locations within an allocentric representation of space. The environmental inputs that underlie the rat's representation of its own location within an environment (in the firing of place cells) are the distances to walls, and different walls are identified by their allocentric direction from the rat. We propose that the locations of goals in an environment is stored downstream of the place cells, in the subiculum. In addition to firing rate coding, place cells may use phase coding relative to the theta rhythm of the EEG. In some circumstances path integration may be used, in addition to environmental information, as an input to the hippocampal system. A detailed computational model of the hippocampus successfully guides the navigation of a mobile robot. The model's behaviour is compared to electrophysiological and behavioural data in rats, and implications for the role of the hippocampus in primates are explored. Burgess, N., Donnett, J.G., Jeffery, K.J. & O'Keefe, J. (1997). Robotic and neuronal simulation of the hippocampus and rat navigation. Review. Philosophical Transactions of the Royal Society of London - Series B: Biological Sciences 352(1360): The properties of hippocampal place cells are reviewed, with particular attention to the nature of the internal and external signals that support their firing. A neuronal simulation of the firing of place cells in open-field environments of varying shape is presented. This simulation is coupled with an existing model of how place-cell firing can be used to drive navigation, and is tested by implementation as a miniature mobile robot. The sensors on the robot provide visual, odometric and short-range proximity data, which are combined to estimate the distance of the walls of the enclosure from the robot and the robot's current heading direction. These inputs drive the hippocampal simulation, in which the robot's location is represented as the firing of place cells. If a goal location is encountered, learning occurs in connections from the concurrently active place cells to a set of 'goal cells', which guide subsequent navigation, allowing the robot to return to an unmarked location. The system shows good agreement with actual placecell firing, and makes predictions regarding the firing of cells in the subiculum, the effect of blocking long-term synaptic changes, and the locus of search of rats after deformation of their environment. Burgess, N., Jackson, A., Hartley, T. & O'Keefe, J. (2000). Predictions derived from modelling the hippocampal role in navigation. Biological Cybernetics Page 9

10 83(3): A computational model of the lesion and single unit data from navigation in rats is reviewed. The model uses external (visual) and internal (odometric) information from the environment to drive the firing of simulated hippocampal place cells. Constraints on the functional form of these inputs are drawn from experiments using an environment of modifiable shape. The place cell representation is used to guide navigation via the creation of a representation of goal location via Hebbian modification of synaptic strengths. The model includes consideration of the phase of firing of place cells with respect to the theta rhythm of hippocampal EEG. A series of predictions for behavioural and single-unit data in rats are derived from the input and output representations of the model. Cho, J., Elgersma, Y., Bombadil, T., Eichenbaum, H., Honavar, V. & Silva, A.J. (2000). Activity of hippocampal CA1 place cells in alpha-camkiit305d mice. Society for Neuroscience Abstracts 30. Place cells are thought to encode spatial information because they fire only when the animal is in specific regions of the environment (place fields). Place fields form rapidly when animals navigate in novel environments, and the spatial selectivity of place cells also improves rapidly as animals explore new environments (Wilson & McNaughton, Science,261,p1055). These phenomena appear to reflect the circuit processes that govern the incorporation of information about a novel environment. Individual place cells tend to show stable place fields upon revisits to familiar environments. The present study investigates the activity of pyramidal cells in region CA1 of mice with a threonine to aspartate mutation at position 305 of the a-calmodulin kinase II (a- CaMKIIT305D). This mutation interferes with calmodulin binding and therefore with kinase activation, and results in severe impairments in a spatial learning task (Morris water maze) and in impaired LTP in the CA1 region of the hippocampus (Elgersma, Fedorov,& Silva, unpublished data). This suggests that the mutation disrupts formation of new memories and/or maintenance and consolidation of memory. Importantly, spatial tasks require learning and incorporation of memory about places. Previous studies reported unstable place fields in a-camkii mutant mice. Although the place cells in these mutant mice showed stability during a recording session, they were unstable across several recording sessions (Cho et al., Science,279,p867; Rotenberg et al., Cell,87,p1351). Supported by: NIH (AG13622) Cowen, S.L., Kudrimoti, H.S., Gerrard, J.L., McNaughton, B.L. & Barnes, C.A. (2000). Three measures of neural ensemble reactivation in the hippocampus fail to reflect reward probabilities present in a familiar task. Society for Neuroscience Abstracts 30. The discovery that waking neural patterns are Page 10

11 reactivated in the hippocampus after behavior presents a possible neural correlate to the memory consolidation process (Wilson and McNaughton, 1994). Consolidation is often affected by level of reward (Kesner et al., 1989). Accordingly, this experiment sought to determine whether hippocampal neural activity associated with high reward locations is preferentially reactivated relative to low reward sites. Three rats were implanted with microdrives, and multiple single-unit activity was recorded from the CA1 layer of the hippocampus. Each rat traversed a T shaped maze during 8 experimental sessions (1 per day). Reward sites were located at the end of each arm and had differential probabilities of containing food (20%, 50%, and 80%). After the fourth session, the location of the 80% and 20% reward sites were exchanged. Maze running periods were preceded and followed by at least 30 minutes of rest. An average of 34 simultaneously recorded units were present during each session. Reactivation during post-exploration periods was assessed using three measures. These measures allowed us to determine whether patterns were reactivated more often and/or with greater intensity. All three measures failed to show a significant relationship between reactivation and reward probability. This result is consistent with the 'incidental' nature of spatial learning and suggests that any enhancement of spatial memory that may result from positive reinforcement is likely to involve encoding processes outside the hippocampus. Supported by: MH46823, AG07434, MH01565, ARCS Cressant, A., Muller, R.U. & Poucet, B. (2002). Remapping of place cell firing patterns after maze rotations. Experimental Brain Research 143(4): When place cells are recorded from rats running on an elevated T-maze inside a curtained enclosure that contains distinct, experimenter selected stimuli, rotations of the maze plus stimuli cause equal rotations of firing fields. Here, we examined the effects of conflicting rotations of a T-maze relative to a laboratory frame that contained a large number of fixed stimuli in the environment and asked whether positional firing patterns stayed in register with the maze or the room cues or were modified in some more complex way. After maze rotations of 90 degrees, 180 degrees or 270 degrees, firing fields were stable in the laboratory frame and thus shifted to a different maze arm. In contrast, rotations of 45 degrees or -45 degrees resulted in dramatic changes of positional firing patterns regardless of their initial position on the maze. Crucially, even cells whose fields were initially on the central platform underwent major firing pattern alterations although the view of the environment from the platform was unchanged by such rotations. Finally, we found that altering the visual appearance by removing without rotation one or two maze arms did not alter firing fields on the remaining part of the maze. Thus, the Page 11

12 "remappings" caused by 45 degrees rotations could result from the disturbed relationship between all arms and the room cues or from the changes in the possible paths the animal can take in the environment. Taken together, our results provide an example of combinatorial coding by the hippocampus, in which the place cell representation of the environment was seen to be modified as a unit and not piecewise according to locally available stimuli. de Araujo, I.E.T., Rolls, E.T. & Stringer, S.M. (2001). A view model which accounts for the spatial fields of hippocampal primate spatial view cells and rat place cells. Hippocampus 11(6): Hippocampal spatial view cells found in primates respond to a region of visual space being looked at, relatively independently of where the monkey is located. Rat place cells have responses which depend on where the rat is located. We investigate the hypothesis that in both types of animal, hippocampal cells respond to a combination of visual cues in the correct spatial relation to each other. In rats, which have a wide visual field, such a combination might define a place. In primates, including humans, which have a much smaller visual field and a fovea which is directed towards a part of the environment, the same mechanism might lead to spatial view cells. A computational model in which the neurons become organized by learning to respond to a combination of a small number of visual cues spread within an angle of a 30 receptive field resulted in cells with visual properties like those of primate spatial view cells. The same model, but operating with a receptive field of 270, produced cells with visual properties like those of rat place cells. Thus a common hippocampal mechanism operating with different visual receptive field sizes could account for some of the visual properties of both place cells in rodents and spatial view cells in primates. Dees, J.A., Terrazas, A., Bohne, K.M., Krause, M., McNaughton, B.L. & Barnes, C.A. (2001). Presence of hippocampal theta during navigation without actual movement. Society for Neuroscience Abstracts 31(643.13): 329. The type-i hippocampal theta rhythm reliably accompanies spatially-directed movements and plays a central role in many computational models of spatial navigation in the mammalian hippocampus. The degree to which self-motion signals such as ambulatory proprioception, optic flow and vestibular inputs drive the theta rhythm is not known. To examine these relationships, rats were trained to drive a car to goal locations on a circular track while differential recordings of the hippocampal EEG were acquired. The car and the platform were driven independently by stepper motors, thus requiring the rat to navigate using primarily optic flow signals when the platform is engaged (WORLD) and optic flow plus vestibular signals when the car is Page 12

13 engaged (CAR). Wavelet analysis was performed on the unfiltered EEG timeseries to determine if any qualitative differences were apparent across spectral bands. In all cases, a prominent 7-9Hz rhythm was easily distinguishable during movement of the CAR and the WORLD and was characteristically reduced during rest. For a subset of sessions (n=4), the animal drove both the CAR and the WORLD on different trials within the same session. Group t-tests comparing the average power in the 7-9Hz band for individual trials in these sessions revealed no significant differences between conditions. These results demonstrate that actual movement is not necessary for the presence of robust hippocampal theta. Supported by: DFG, AG12609 & MH01565 Eichenbaum, H., Dudchenko, P., Wood, E., Shapiro, M. & Tanila, H. (1999). The hippocampus, memory, and place cells: Is it spatial memory or a memory space? Neuron 23( ). Foster, D.J., Morris, R.G.M. & Dayan, P. (2000). A model of hippocampally dependent navigation, using the temporal difference learning rule. Hippocampus 10(1): This paper presents a model of how hippocampal place cells might be used for spatial navigation in two watermaze tasks: the standard reference memory task and a delayed matching-to-place task. In the reference memory task, the escape platform occupies a single location and rats gradually learn relatively direct paths to the goal over the course of days, in each of which they perform a fixed number of trials. In the delayed matching-to-place task, the escape platform occupies a novel location on each day, and rats gradually acquire one-trial learning, i.e., direct paths on the second trial of each day. The model uses a local, incremental, and statistically efficient connectionist algorithm called temporal difference learning in two distinct components. The first is a reinforcement-based actor-critic network that is a general model of classical and instrumental conditioning. In this case, it is applied to navigation, using place cells to provide information about state. By itself, the actor-critic can learn the reference memory task, but this learning is inflexible to changes to the platform location. We argue that one-trial learning in the delayed matching-to-place task demands a goal-independent representation of space. This is provided by the second component of the model: a network that uses temporal difference learning and self-motion information to acquire consistent spatial coordinates in the environment. Each component of the model is necessary at a different stage of the task; the actor-critic provides a way of transferring control to the component that performs best. The model successfully captures gradual acquisition in both tasks, and, in particular, the ultimate development of one-trial Page 13

14 learning in the delayed matching-to-place task. Place cells report a form of stable, allocentric information that is well-suited to the various kinds of learning in the model. Frank, L.M., Brown, E.N. & Wilson, M.A. (2000). Trajectory encoding in the hippocampus and entorhinal cortex. Neuron 27: We recorded from single neurons in the hippocampus and entorhinal cortex (EC) of rats to investigate the role of these structures in navigation and memory representation. Our results revealed two novel phenomena: first, many cells in CA1 and the EC fired at significantly different rates when the animal was in the same position depending on where the animal had come from or where it was going. Second, cells in deep layers of the EC, the targets of hippocampal outputs, appeared to represent the similarities between locations on spatially distinct trajectories through the environment. Our findings suggest that the hippocampus represents the animal's position in the context of a trajectory through space and that the EC represents regularities across different trajectories that could allow for generalization across experiences. Frank, L.M., Brown, E.N. & Wilson, M.A. (2001). A comparison of the firing properties of putative excitatory and inhibitory neurons from CA1 and the entorhinal cortex. Journal of Neurophysiology 86(4): The superficial layers of the entorhinal cortex (EC) provide the majority of the neocortical input to the hippocampus, and the deep layers of the EC receive the majority of neocortically bound hippocampal outputs. To characterize information transmission through the hippocampal and EC circuitry, we recorded simultaneously from neurons in the superficial EC, the CA1 region of hippocampus, and the deep EC while rodents ran for food reward in two environments. Spike waveform analysis allowed us to classify units as fast-spiking (FS) putative inhibitory cells or putative excitatory (PE) cells. PE and FS units' firing were often strongly correlated at short time scales, suggesting the presence a monosynaptic connection from the PE to FS units. EC PE units, unlike those found in CA1, showed little or no tendency to fire in bursts. We also found that the firing of FS and PE units from all regions was modulated by the approximately 8 Hz theta rhythm, although the firing of deep EC FS units tended to be less strongly modulated than that of the other types of units. When we examined the spatial specificity of FS units, we determined that FS units in all three regions showed low specificity. At the same time, retrospective coding, in which firing rates were related to past position, was present in FS units from all three regions and deep EC FS units often fired in a Page 14

15 "path equivalent" manner in that they were active in physically different, but behaviorally related positions both within and across environments. Our results suggest that while the firing of FS units from CA1 and the EC show similarly low levels of position specificity, FS units from each region differ from one another in that they mirrored the associated PE units in terms of their tendency to show more complex positional firing properties like retrospective coding and path equivalence. Frank, L.M., Eden, U.T., Solo, V., Wilson, M.A. & Brown, E.N. (2002). Contrasting patterns of receptive field plasticity in the hippocampus and the entorhinal cortex: An adaptive filtering approach. Journal of Neuroscience 22(9): Neural receptive fields are frequently plastic: a neural response to a stimulus can change over time as a result of experience. We developed an adaptive point process filtering algorithm that allowed us to estimate the dynamics of both the spatial receptive field (spatial intensity function) and the interspike interval structure (temporal intensity function) of neural spike trains on a millisecond time scale without binning over time or space. We applied this algorithm to both simulated data and recordings of putative excitatory neurons from the CA1 region of the hippocampus and the deep layers of the entorhinal cortex (EC) of awake, behaving rats. Our simulation results demonstrate that the algorithm accurately tracks simultaneous changes in the spatial and temporal structure of the spike train. When we applied the algorithm to experimental data, we found consistent patterns of plasticity in the spatial and temporal intensity functions of both CA1 and deep EC neurons. These patterns tended to be opposite in sign, in that the spatial intensity functions of CA1 neurons showed a consistent increase over time, whereas those of deep EC neurons tended to decrease, and the temporal intensity functions of CA1 neurons showed a consistent increase only in the "theta" ( msec) region, whereas those of deep EC neurons decreased in the region between 20 and 75 msec. In addition, the minority of deep EC neurons whose spatial intensity functions increased in area over time fired in a significantly more spatially specific manner than non-increasing deep EC neurons. We hypothesize that this subset of deep EC neurons may receive more direct input from CA1 and may be part of a neural circuit that transmits information about the animal's location to the neocortex. Fuhs, M.C., Skaggs, W.E. & Touretzky, D.S. (2000). Modeling experiencedependent remapping in rat hippocampus. Society for Neuroscience Abstracts 30. Small environmental changes can cause radical rearrangement of hippocampal place fields ("remapping"). In some cases remapping occurs abruptly, but only after several exposures (Bostock et al. 1990), implying that remapping may depend on Page 15

16 learning. We have implemented a neural network model of this phenomenon based on attractor dynamics and Hebbian learning. The model contains three groups of cells, representing entorhinal cortex, dentate gyrus granule cells, and a group of inhibitory interneurons. The granule cells and interneurons are assumed to inhibit each other via modifiable connections, and to receive excitatory input from the entorhinal cells. Recurrent connections among granule cells give rise to multiple stable attractor basins. In an unfamiliar environment, strengthening of connections from granule to inhibitory cells counteracts the entorhinal input to the inhibitory cells, eventually suppressing inhibitory cell activity. Once the environment becomes familiar, the granule-to-interneuron connections cease to be modifiable. If small changes are then made to the environment, the resulting change in EC activity permits a small subset of interneurons to become active. When this happens repeatedly, the active interneuron-to-granule-cell connections potentiate, eventually resulting in destabilization of the granule cell attractor basin. The system then settles into a new basin and begins learning the features of this "new" environment. The model predicts that the probability of remapping as a function of exposures should be bimodally distributed, and that a subset of DG interneurons should show elevated activity in novel environments.supported by: NSF and NSF graduate fellowship Fuhs, M.C., Touretzky, D.S. & Skaggs, W.E. (2001). Do hippocampal place cell ensembles behave coherently in stretched environments? Society for Neuroscience Abstracts 31(643.11): 329. O'Keefe and Burgess (1996) measured hippocampal place fields in a rectangular environment whose walls could be stretched. They found fields altered in a way that resembled a sum of up to four Gaussians, each tuned to the distance from the field center to one of the walls. Stretching shifts the Gaussians relative to each other, changing the field shapes and the relationships between fields. The current experiment examined whether place cell activity could be better explained by a rigid map whose binding to the environment shifts as the rat moves in the stretched arena, thus producing the observed place field distortions. Data were recorded simultaneously from multiple CA1 pyramidal cells in both a control session and multiple stretched sessions, using a 12 tetrode recording system. Each neuron's stretched session activity at each timestep was estimated by using the ensemble activity pattern of the other cells to compute a similarityweighted average of the neuron's control session activity. If the map is rigid, this ensemble prediction should be accurate. In a data set with 25 cells, the ensemble method did slightly better than the sum of Gaussians in predicting cell activity, but the difference was not statistically significant. Thus, the test was inconclusive. Page 16

17 However, the comparison was impaired by a relatively low yield of simultaneously recorded cells. Future data will test whether the performance of the ensemble method improves when ensembles containing larger numbers of neurons are used. Supported by: NSF Graduate Research Fellowship (Fuhs) and NSF DGE Gaussier, P., Revel, A., Banquet, J.P. & Babeau, V. (2002). From view cells and place cells to cognitive map learning: processing stages of the hippocampal system. Biological Cybernetics 86(1): The goal of this paper is to propose a model of the hippocampal system that reconciles the presence of neurons that look like "place cells" with the implication of the hippocampus (Hs) in other cognitive tasks (e.g., complex conditioning acquisition and memory tasks). In the proposed model, "place cells" or "view cells" are learned in the perirhinal and entorhinal cortex. The role of the Hs is not fundamentally dedicated to navigation or map building, the Hs is used to learn, store, and predict transitions between multimodal states. This transition prediction mechanism could be important for novelty detection but, above all, it is crucial to merge planning and sensory-motor functions in a single and coherent system. A neural architecture embedding this model has been successfully tested on an autonomous robot, during navigation and planning in an open environment. Gerrard, J.L., Bower, M.R., Insel, N., Lipa, P., Barnes, C.A. & McNaughton, B.L. (2001). A long day's journey into night. Society for Neuroscience Abstracts 31(643.12): 329. Are patterns of coactivity of hippocampal neurons selected at random, or is there a propensity for some cells to exhibit strong correlations over many different experiences? This question has been the subject of controversy for several decades. Most studies have been performed in one or a few rather small environments in which the sample of the possible "state space" of the network is small. CA1 pyramidal cell activity was recorded while rats walked down and back along a long (13m x 2m) corridor. Few locations were visited more than once. As predicted by the random allocation model, the variance of the distribution of firing rate correlations became compressed around zero as compared to typical periods of repetitive track running. Moreover, the distribution of mean firing rates became significantly less sparse, again in agreement with the random model. This suggests that synaptic weight vectors of CA1 pyramidal cells are essentially uncorrelated. The degree of memory trace reactivation (Wilson and McNaughton, Science, 265: , 1994; Kudrimoti et al., J. Neurosci., 19: , 1999) was also assessed during sleep following the long corridor experience, and compared to rats which ran repetitively around a small track. The population Page 17

18 vector overlap between behavior and subsequent sleep, and the percent of the firing rate correlation variance during sleep that was statistically explained by the pattern during behavior (robust measures of reactivation) were significantly lower in the long corridor experiment than for the typical case of repetitious behavior, suggesting that repetition facilitates memory trace reactivation. Supported by: AG12609, MH01565, MH46823 & ARCS Gothard, K.M., Skaggs, W.E. & McNaughton, B.L. (1996). Dynamics of mismatch correction in the hippocampal ensemble code for space: Interaction between path integration and environmental cues. Journal of Neuroscience 16: Populations of hippocampal neurons were recorded simultaneously in rats shuttling on a track between a fixed reward site at one end and a movable reward site, mounted in a sliding box, at the opposite end. While the rat ran toward the fixed site, the box was moved. The rat returned to the box in its new position. On the initial part of all journeys, cells fired at fixed distances from the origin, whereas on the final part, cells fired at fixed distances from the destination. Thus, on outward journeys from the box, with the box behind the rat, the position representation must have been updated by path integration. Farther along the journey, the place field map became aligned on the basis of external stimuli. The spatial representation was quantified in terms of population vectors. During shortened journeys, the vector shifted from an alignment with the origin to an alignment with the destination. The dynamics depended on the degree of mismatch with respect to the full-length journey. For small mismatches, the vector moved smoothly through intervening coordinates until the mismatch was corrected. For large mismatches, it jumped abruptly to the new coordinate. Thus, when mismatches occur, path integration and external cues interact competitively to control placecell firing. When the same box was used in a different environment, it controlled the alignment of a different set of place cells. These data suggest that although map alignment can be controlled by landmarks, hippocampal neurons do not explicitly represent objects or events. Gothard, K.M., Skaggs, W.E., Moore, K.M. & McNaughton, B.L. (1996). Binding of hippocampal CA1 neural activity to multiple reference frames in a landmark-based navigation task. Journal of Neuroscience 16: The behavioral correlates of rat hippocampal CA1 cells were examined in a spatial navigation task in which two cylindrical landmarks predicted the location of food. The landmarks were maintained at a constant distance from each other but were moved from trial to trial within a large arena surrounded by static background cues. Page 18