The hippocampus as a spatial map — Dostrovsky & O'Keefe (Brain Research, 1971)

Experimental context and core hypothesis

• Hippocampus thought to function as a spatial map; initial observations show rats with hippocampal damage are hyperactive in novel environments, perseverative, resistant to extinction on learned tasks, and poor at spatial tasks (e.g., mazes) or tasks requiring response changes across trials. These deficits suggested loss of a neural system that provides a cognitive, spatial map of the environment.
• This paper provides preliminary electrophysiological support for the spatial-map theory by recording hippocampal unit activity in freely moving rats.
• Context: previous work (e.g., Vanderwolf) linked certain behaviors to hippocampal EEG patterns; this study seeks cellular correlates of spatial processing during natural behavior.

Hardware and recording apparatus

• Recording system: a small, lightweight microdrive carrying up to $8$ glass-insulated platinum-plated tungsten microelectrodes; permanently fixed to the rat’s skull.
• Each electrode can be moved independently via set screws on the manipulator.
• Signal quality improvement: two adjacent microelectrodes feed a high-input-impedance differential FET preamplifier mounted on the microdrive to reject muscle and movement artifacts; signals then travel through flexible, lightly screened wires to standard recording equipment.
• Early approach: two electrodes glued side by side with tips separated by about $0.5 ext{ mm}$ and advanced together through the hippocampus.
• Later approach: one electrode placed in cortical white matter to serve as a reference for other electrodes.
• Recording sites are locally identified by ensuring which electrode carries the unit during sessions.

Surgical procedures and postoperative protocol

• Anesthesia and fixation: rats anesthetized with Equithesin and positioned in a stereotaxic instrument.
• Implantation: microdrive assembly fixed to the skull with dental cement so electrode tips pass through a skull hole and rest in the upper cortical layers.
• Postoperative recovery: after recovery from anesthesia, recordings begin once the animal is comfortable; the first electrode or pair is gradually advanced through the cortex into the dorsal hippocampus to search for units.
• Localizing recording position: recording sites are left in final positions for at least $24 ext{ h}$ to allow gliosis around tips; animals are then sacrificed, perfused with $10\%\,formol{-}saline$, and brains sectioned and stained (cresyl violet or Darrow red with Luxol fast blue).
• Recording site localization: each site is calculated as the distance on the microdrive above the final position after histology.

Recording stand and behavioral sessions

• Recording stand: a raised platform of $24\text{ cm} \times 36\text{ cm}$, surrounded on three sides by a white plastic curtain; the fourth side is open to give the rat a view of the laboratory.
• Behavioral repertoire during recordings:
  • Spontaneous: walking, eating, drinking, grooming, sleeping.
  • Elicited: orienting, sniffing at cotton wool or various odours, biting a polyethylene tube, bar-pressing for food.
  • Stimulus testing: simple auditory (clicks, whistles, scratching noises), visual (moving light, hand, striped board), olfactory (odours, rat faeces), and tactile (touch and pressure on the body surface).
• Duration of unit observation: each unit studied for at least $15\text{ min}$, most for more than $30\text{ min}$.

Recording sites and unit sample

• Sample: among the data from $36$ electrode penetrations through the dorsal hippocampus (fields CA1 and CA4, and dentate gyrus) in $23$ rats, recording yielded $76$ units.
• Unit breakdown:
  • $8$ units were identified as potentially spatially specific (see below).
  • $14$ units classified as arousal/attentional.
  • $21$ units categorized as movement-related, with firing linked to some behaviors (orienting, sniffing, bar-pressing, walking) and little to no firing during eating, drinking, grooming, quiet sitting, or slow-wave sleep.
  • $2$ units with properties related to the animal’s expectations.
  • Remaining $31$ units showed no adequate stimulus, behavior, or consistent responses; within this group, $15$ were essentially silent most of the time, except occasional spikes in conjunction with bursts from multiple smaller units.
• Interpretation: the key finding rests with the subset of $8$ units whose activity was tied to spatial position and orientation, forming the basis for the spatial-map hypothesis.

Spatial orientation units: characteristics

• The $8$ orientation-preferring units fired solely or maximally when the rat was located in a particular part of the testing platform and facing a specific direction.
• Among these, $5$ units were similar to the silent group in other respects: they did not fire unless the animal was moderately aroused, in the correct platform location, and, for $4\overline{5}$ of them, receiving the appropriate sensory stimulus.
• The remaining $3$ units possessed a spontaneous firing rate that was maximal in a particular portion of the testing stand.
• Conclusion from this cluster: these cells showed a preferred spatial orientation, i.e., directional selectivity linked to the animal’s position.

Figure 1: Example unit in CA1

• An example unit was recorded in the CA1 field of the anterior dorsal hippocampus.
• Properties:
  • The unit had no spontaneous activity and fired only when the rat faced directions marked A or B and was simultaneously lightly restrained by hand over the back/shoulder.
  • Both the spatial orientation (A or B) and the tactile stimulus were necessary for activation.
• Experimental procedure (as illustrated in Fig. 1): the rat was moved counterclockwise around the platform, starting/ending at position A. At each labeled position, the rat was held for a few seconds, then released and moved to the next position.
• Observations:
  • The unit fired only when the rat was restrained at A or B and during the movement from B to the next position, while arousal remained high.
  • No firing occurred between A and B when the rat was unrestrained in the same location, demonstrating the requirement of tactile stimulation for activation.
  • Other tactile strategies (e.g., loosely holding around the trunk) did not produce a response.
• Orientation cues:
  • For this unit, spatial orientation cues appeared to be visual because turning off the lights abolished the response.
  • The spatial orientation field size was typical for all observed units.
• Additional unit behavior:
  • A different unit, located just above the CA1 unit, fired in response to any moving visual stimulus when the rat was between positions G and A. When moderately aroused, this unit fired only at this orientation and was silent elsewhere; at higher arousal, weak responses could be elicited at other positions, but the preferred-direction response remained strong.

Cues and sensory cues tested

• Multiple strategies were employed to identify cues responsible for orientation preferences:
  • Rotating the platform to test proximal olfactory or tactile cues.
  • Shutting off prominent noise sources (fans, polygraph machines).
  • Darkening or occluding parts of the visual field.
• Outcome:
  • None of these manipulations yielded notably successful disruption of orientation preferences.
  • In some cases, removing the curtain or radical environmental changes disrupted the orientation, but the rats then explored as if in a totally new environment, suggesting the cues are equipotential rather than single, isolated signals.
• Implication: the orientation preferences likely arise from several interchangeable cues; removing any one cue is not sufficient to abolish the response.

Stability, environment perturbations, and map concept

• Key interpretation: the hippocampus provides a spatial reference map to the brain, indicating the rat’s direction relative to landmarks and the occurrence of sensory events while facing a particular orientation.
• The proposed internal wiring:
  • Activation of cells specifying a given orientation, together with movement-related signals (hippocampal movement-related units), would tend to activate cells representing adjacent or subsequent spatial orientations.
  • This wiring would allow the map to anticipate sensory consequences of a move, i.e., sensory events associated with a particular trajectory.
• Temporal anticipation (context from Vinogradova): there is evidence of anticipation in the temporal domain by hippocampal cells; two units in this study also showed behavior tied to expectation.
• Example of anticipation: a unit in the dentate gyrus fired when the rat sniffed its own water dish; when the dish was covered by a beaker, the response occurred the first time the rat returned to sniff, attenuated on the second sniff, and absent thereafter until the beaker was removed.
• Implications of mismatch: mismatches between anticipated and experienced stimuli would trigger exploration to incorporate into the map new stimulus patterns.
• Consequences of map loss: without the map, rats would not immediately notice environmental changes and would rely on rigid, less flexible strategies to reach places (e.g., turning at junctions, following odors, avoiding bright lights).
• Synthesis: loss of the spatial reference map could account for many of the behavioral deficits observed after hippocampal lesions.

Tolman, cognitive maps, and real-world significance

• Tolman’s notion of cognitive maps in animals and humans provides a theoretical anchor for interpreting the hippocampus as a spatial map.
• The present results extend this idea by offering a cellular basis for a cognitive map: specific hippocampal neurons encode spatial orientation and, in conjunction with movement signals, could generate a dynamic map that predicts sensory outcomes of movement.
• Practical implications:
  • A neural substrate for flexible navigation and adaptation to environmental changes.
  • Potential explanations for deficits in spatial learning and navigation observed after hippocampal damage.
  • Supports the view that spatial cognition relies on an internal framework rather than solely on stimulus–response associations.

The broader interpretation and theoretical implications

• The authors propose that hippocampal cells forming a spatial map would provide directional information relative to landmarks and expected sensory inputs, integrated with movement signals to form a predictive representation of space.
• The map could enable planning of routes by predicting sensory consequences of different trajectories, aligning with broader theories of predictive coding in neural systems.
• The finding that some orientation-selective units require multiple cues (vision, tactile input, arousal) supports a distributed, multi-sensory basis for spatial representation rather than a single, modality-specific cue.
• The observation that altering environmental cues (like removing curtains) sometimes disrupts orientation but not permanently suggests an adaptable, cue-composite map whose components can be re-weighted by experience.

Ethical, philosophical, and practical implications

• Ethically: animal research here involves invasive recording in awake, behaving animals; the study was conducted with appropriate care and reported support from the Medical Research Council and collaboration acknowledgments.
• Philosophical: provides empirical support for cognitive maps as neural constructs, reinforcing the view that memory and navigation are grounded in brain-wide representations rather than purely behaviorist associations.
• Practical: insights into spatial memory circuits could inform understanding of human spatial memory disorders and inform computational models of navigation in artificial systems (e.g., robotics) that simulate place- and route-based representations.

Key numerical references and methodological details (summary)

• Equipment capacity: up to $8$ electrodes on a microdrive.
• Initial electrode spacing: tip-to-tip separation of $0.5\ \text{mm}$.
• Recording platform dimensions: $24\text{ cm} \times 36\text{ cm}$.
• Recording duration per unit: at least $15\min$ (most >$30\min$).
• Post-mortem processing: brains perfused with $10\%\text{ formol-saline}$; histology with cresyl violet or Darrow red + Luxol fast blue.
• Number of rats: $23$.
• Electrode penetrations: $36$ through dorsal hippocampus and dentate gyrus.
• Unit yield: $76$ units recorded; focus on $8$ units with spatial orientation properties.
• Time calibration shown in Fig. 1 data: $400\text{ ms}$.
• Anatomical detail: one CA1 unit discussed in anterior dorsal hippocampus; other units recorded in the vicinity (CA1, CA3/CA4 designation in text).

Acknowledgments and references (contextual)

• Acknowledgments to L. Nadel, G. D. Gaffan, P. D. Wall for discussion and criticism; technical assistance from E. G. Merrill, A. Ainsworth, R. Sampson, J. Astafiev, J. Sheppard.
• Funding: Medical Research Council; J. Dostrovsky noted as Edmund Davis Scholar.
• Related references cited include foundational works by Tolman on cognitive maps, Vanderwolf on hippocampal EEG with movement, Vinogradova on anticipatory hippocampal activity, and earlier recording techniques in awake, freely moving animals.

Takeaways for exam preparation

• The hippocampus contains neurons with spatially selective responses tied to the animal’s orientation and position, supporting the view of a neural spatial map.
• Spatially selective units require combination of orientation, position, and sensory cues; no single cue is solely responsible.
• The map is predictive: activity in orientation-specific cells, together with movement signals, could anticipate sensory outcomes of movement, enabling flexible navigation.
• Disruption of the map leads to navigation deficits and insensitivity to environmental changes, providing a neural account for some behavioral deficits seen after hippocampal damage.
• The work links cellular activity to Tolman’s cognitive maps and sets the stage for understanding how memory and spatial navigation are integrated in the hippocampus.