Periodic Reporting for period 4 - NavigationCircuits (Neural circuits for route planning in goal-directed spatial navigation)
Reporting period: 2021-11-01 to 2023-03-31
Spatial navigation is an essential ability for animals to survive in a geometric space. A prominent feature of rats is their ability to create an internal metric of space, or a cognitive map, in their brain. While place cells in the hippocampus are considered key elements of the spatial representation system in the brain, the activity of these cells primarily depends on the animal’s own position. It is thus not clear how the brain computes an estimate of future positions, necessary for route planning. The main objective of this project is to clarify the wider brain circuits necessary for goal decision and route planning. A growing body of evidence indicates key roles for the prefrontal cortex (PFC) and the retrosplenial cortex (RSC) in navigation. We hypothesised that RSC, downstream of hippocampal area CA1, may represent the animal’s future position by making use of information from place cells about positions and movements. The future positions may then be evaluated in PFC, a downstream target of RSC and CA1, which potentially represents the spatial proximity to a goal.
We investigated this question by using the state-of-the-art high-density recording of neural activity from navigating rats. The project’s achievements are roughly summarised in two directions. First, we identified that the medial entorhinal cortex (MEC) is the major source of spatial information in the RSC, revealing a pathway involved in the transformation of the brain’s spatial map representation into the one that is referenced to the animal’s own perspective. Another major achievement is the discovery of the brain’s goal map in a subregion of the PFC, the orbitofrontal cortex (OFC). We found that neurons in the OFC represent the animal’s subsequent destination throughout the journey and this representation can even predict the animal’s incorrect destination on an error trial.
These results together defined navigation circuits in the brain from a wider perspective beyond previously well-studied hippocampal formation. Our findings will inspire further investigations of the mechanism for the brain to plan a future navigation journey in a complex environment, and may also provide insights into the pathophysiology of navigation deficits in patients with dementia or Alzheimer’s disease.
We investigated this question by using the state-of-the-art high-density recording of neural activity from navigating rats. The project’s achievements are roughly summarised in two directions. First, we identified that the medial entorhinal cortex (MEC) is the major source of spatial information in the RSC, revealing a pathway involved in the transformation of the brain’s spatial map representation into the one that is referenced to the animal’s own perspective. Another major achievement is the discovery of the brain’s goal map in a subregion of the PFC, the orbitofrontal cortex (OFC). We found that neurons in the OFC represent the animal’s subsequent destination throughout the journey and this representation can even predict the animal’s incorrect destination on an error trial.
These results together defined navigation circuits in the brain from a wider perspective beyond previously well-studied hippocampal formation. Our findings will inspire further investigations of the mechanism for the brain to plan a future navigation journey in a complex environment, and may also provide insights into the pathophysiology of navigation deficits in patients with dementia or Alzheimer’s disease.
At the beginning of the project, we focused on several methodological and technological developments that are necessary to achieve the project's aims. We constructed new behavioural mazes, designed our own electrode recording device, and then established a strategy to combine high-density recordings together with optogenetic circuit manipulations from behaving rats. By using these methodologies, we investigated a scientific question proposed in each Work Package, as described in the following:
The primary aim of WP1 is to understand how an animal plans an efficient route to a destination by avoiding obstacles along the journey. For this aim, we have designed a modified version of the goal-directed navigation task in an open-field arena (Pfeiffer and Foster, 2013). The major modifications in our task design can be summarised in the following two points. First, we introduced a wall in the maze, so that the animal is required to find a goal-directed path avoiding the wall. Second, we let the animals perform this task in complete darkness, imposing the animals to rely on a cognitive map, rather than sensory perceptions. We confirmed that rats could successfully learn the position of the wall and take a smooth wall-avoiding path to the destination, supporting that rats use a cognitive map, especially under limited access to sensory signals. As a neural mechanism supporting this ability, we discovered that neurons in the hippocampus exhibit firing before starting the navigation, which corresponds to the animal’s subsequent goal-directed journey. We then asked whether this goal-directed activity in the hippocampus depends on the goal information provided by the prefrontal cortex. We inactivated neurons in the nucleus reuniens - an anatomical hub between the prefrontal cortex and the hippocampus - and found that the animals exhibited a significant impairment in taking an efficient route to the goal. Corresponding to this deficit, we also found that the goal-directed activity in the hippocampus was largely diminished during the silencing of reuniens neurons. Together, the WP1 identified a neural circuit that plays a key role in route planning for navigation.
The main achievement of WP2 is the discovery of a goal map in a subregion of the prefrontal cortex - the orbitofrontal cortex (Basu et al., Nature 2021). This finding inspired further investigation of its circuit mechanisms. For example, we are currently investigating two questions; 1) how this goal information can influence the goal-directed activity in the hippocampus, and 2) whether or not a goal decision in the orbitofrontal cortex is guided by the dopamine-striatum system.
In WP3, we investigated a neural representation of flexible route decisions. For this aim, we designed a new behavioural task in which a rat is forced to take different paths to reach the same goal locations. This task design allowed us to investigate whether the prefrontal cortex forms a goal representation irrespective of route choices, dissociating a representation of goals from that for routes or actions. By using this behavioural task, we confirmed that neurons in the orbitofrontal cortex form a goal representation that is consistent irrespective of route choices, suggesting that the OFC’s goal representation is in a rather abstract fashion.
Finally, in WP4, we had to change our initial plan because, unlike my original hypothesis, we found that the hippocampal CA1 neurons do not strongly project to the retrosplenial cortex (RSC). We thus instead focused on the connections from the medial entorhinal cortex (MEC) to the RSC and discovered that the information about environmental boundaries in the MEC is transferred to the RSC but with a different coordinate system from a world-centred allocentric to self-referenced egocentric coordinate frame (van Wijngaarden et al., eLife 2020).
Altogether, the project successfully identified the wider brain circuits necessary for navigation beyond well-studied hippocampal formation. As spatial navigation requires multistep computations, our results provided a key step toward a holistic understanding of navigation circuits in the brain.
The primary aim of WP1 is to understand how an animal plans an efficient route to a destination by avoiding obstacles along the journey. For this aim, we have designed a modified version of the goal-directed navigation task in an open-field arena (Pfeiffer and Foster, 2013). The major modifications in our task design can be summarised in the following two points. First, we introduced a wall in the maze, so that the animal is required to find a goal-directed path avoiding the wall. Second, we let the animals perform this task in complete darkness, imposing the animals to rely on a cognitive map, rather than sensory perceptions. We confirmed that rats could successfully learn the position of the wall and take a smooth wall-avoiding path to the destination, supporting that rats use a cognitive map, especially under limited access to sensory signals. As a neural mechanism supporting this ability, we discovered that neurons in the hippocampus exhibit firing before starting the navigation, which corresponds to the animal’s subsequent goal-directed journey. We then asked whether this goal-directed activity in the hippocampus depends on the goal information provided by the prefrontal cortex. We inactivated neurons in the nucleus reuniens - an anatomical hub between the prefrontal cortex and the hippocampus - and found that the animals exhibited a significant impairment in taking an efficient route to the goal. Corresponding to this deficit, we also found that the goal-directed activity in the hippocampus was largely diminished during the silencing of reuniens neurons. Together, the WP1 identified a neural circuit that plays a key role in route planning for navigation.
The main achievement of WP2 is the discovery of a goal map in a subregion of the prefrontal cortex - the orbitofrontal cortex (Basu et al., Nature 2021). This finding inspired further investigation of its circuit mechanisms. For example, we are currently investigating two questions; 1) how this goal information can influence the goal-directed activity in the hippocampus, and 2) whether or not a goal decision in the orbitofrontal cortex is guided by the dopamine-striatum system.
In WP3, we investigated a neural representation of flexible route decisions. For this aim, we designed a new behavioural task in which a rat is forced to take different paths to reach the same goal locations. This task design allowed us to investigate whether the prefrontal cortex forms a goal representation irrespective of route choices, dissociating a representation of goals from that for routes or actions. By using this behavioural task, we confirmed that neurons in the orbitofrontal cortex form a goal representation that is consistent irrespective of route choices, suggesting that the OFC’s goal representation is in a rather abstract fashion.
Finally, in WP4, we had to change our initial plan because, unlike my original hypothesis, we found that the hippocampal CA1 neurons do not strongly project to the retrosplenial cortex (RSC). We thus instead focused on the connections from the medial entorhinal cortex (MEC) to the RSC and discovered that the information about environmental boundaries in the MEC is transferred to the RSC but with a different coordinate system from a world-centred allocentric to self-referenced egocentric coordinate frame (van Wijngaarden et al., eLife 2020).
Altogether, the project successfully identified the wider brain circuits necessary for navigation beyond well-studied hippocampal formation. As spatial navigation requires multistep computations, our results provided a key step toward a holistic understanding of navigation circuits in the brain.
We introduced a miniature microscope by Inscopix, and successfully implemented the calcium imaging of neurons in the retrosplenial cortex (RSC) and in the hippocampus. The major advantage of this method is a long-term monitoring of the same neurons. We are currently using this technique to understand the development of goal-selective activity in hippocampal place cells.
An unexpected discovery was made when we recorded from RSC neurons while rats performed random foraging in an open-field square box. We found a subpopulation of neurons in RSC encode environmental boundaries, and furthermore, demonstrated that RSC can implement a transformation from an allocentric (world-centred) to egocentric (self-centred) coordinate system. This finding has been published in eLife (van Wijngaarden et al., eLife 2020).
An unexpected discovery was made when we recorded from RSC neurons while rats performed random foraging in an open-field square box. We found a subpopulation of neurons in RSC encode environmental boundaries, and furthermore, demonstrated that RSC can implement a transformation from an allocentric (world-centred) to egocentric (self-centred) coordinate system. This finding has been published in eLife (van Wijngaarden et al., eLife 2020).