Neural Networks for Navigation: From Connections to Computations

Abstract

Many animals can navigate toward a goal they cannot see based on an internal representation of that goal in the brain's spatial maps. These maps are organized around networks with stable fixed-point dynamics (attractors), anchored to landmarks, and reciprocally connected to motor control. This review summarizes recent progress in understanding these networks, focusing on studies in arthropods. One factor driving recent progress is the availability of the Drosophila connectome; however, it is increasingly clear that navigation depends on ongoing synaptic plasticity in these networks. Functional synapses appear to be continually reselected from the set of anatomical potential synapses based on the interaction of Hebbian learning rules, sensory feedback, attractor dynamics, and neuromodulation. This can explain how the brain's maps of space are rapidly updated; it may also explain how the brain can initialize goals as stable fixed points for navigation.

Fine-grained descending control of steering in walkingDrosophila

“…The links between navigation and locomotor control are unclear in vertebrates, but these links have recently been described in arthropods 62,63,75 . Given that brain circuits for navigation have clear homologies in arthropods and vertebrates 99,100 , it seems likely that insights from arthropods will have general relevance.…”

Encoding of interdependent features of head direction and angular head velocity in navigation

“…Interestingly, both SP and MP neurons selective to HD have been discovered across various brain regions, including the anterior dorsal nucleus (ADn), Postsubiculum (PoS), retrosplenial cortex, and entorhinal cortex (Hennestad et al, 2021; Taube, 2007), with focus on SP neurons possibly for their majority and simplicity in HD selectivity and ease in computational modeling (Johnson et al, 2005; Wilson, 2023). Inspired by the aforementioned computational modeling, here we delved into the neural representations of HD and AHV in the PoS, the primary cortex of the HD system in mice (Clark and Taube, 2012; Sharp et al, 2001; Winter and Taube, 2014), where neurons attune to both HD and AHV (Sharp, 1996).…”

“…Interestingly, both SP and MP neurons selective to HD have been discovered across various brain regions, including the anterior dorsal nucleus (ADn), Postsubiculum (PoS), retrosplenial cortex, and entorhinal cortex (Hennestad et al, 2021;Taube, 2007), with focus on SP neurons possibly for their majority and simplicity in HD selectivity and ease in computational modeling (Johnson et al, 2005;Wilson, 2023).…”

A rotational velocity estimate constructed through visuomotor competition updates the fly’s neural compass

“…Directional sensory information is conveyed to the compass by distinct populations of 'ring neurons' that are tuned to a variety of cues, such as polarized light (Hardcastle et al, 2020), prominent visual landmarks (Seelig and Jayaraman, 2013), and the wind (Okubo et al, 2020). Through anti-Hebbian plasticity, ring neurons construct a map that tethers the HD bump's position to the orientation of these directional sensory signals (Fisher et al, 2019;Fisher et al, 2022;Kim, 2021;Kim et al, 2019;Wilson, 2023). An important feature of this plasticity processes is that the ring neuron map should be strongest in the presence of highly reliable directional sensory information (Haberkern et al, 2022;Kim et al, 2019).…”

“…Recent work leveraging the powerful circuit tools available in Drosophila has begun to provide a circuit-and cell-type level understanding of sensorimotor-driven navigation computations, setting the stage for a deeper understanding of how self-motion estimates are constructed and ultimately used to update internal representations Green et al, 2017;Hulse et al, 2021;Lu et al, 2022;Lyu et al, 2022;Matheson et al, 2022;Shiozaki et al, 2020;Stone et al, 2017;Turner-Evans et al, 2017). These navigation computations are housed in a highly conserved brain region known as the central complex whose intricately woven circuits implement a variety of vector-based navigation computations (Fisher, 2022;Honkanen et al, 2019;Pfeiffer and Homberg, 2014;Turner-Evans and Jayaraman, 2016;Wilson, 2023), beginning with the computation of HD in the EB (Figure 1A (Hulse and Jayaraman, 2019;Seelig and Jayaraman, 2015)). Flies update their HD representation by continuously integrating rotational velocity information and by relying on directional sensory cues like visual landmarks when present (Fisher et al, 2019;Fisher et al, 2022;Haberkern et al, 2022;Kim, 2021;Kim et al, 2019;Okubo et al, 2020;Seelig and Jayaraman, 2013;Sun et al, 2017).…”