Adaptive Motor Control RIKEN Hakubi Research Team 

Fujiwara Lab

My lab’s goal is to reveal the neural bases of motor control at unprecedented spatiotemporal resolution using a powerful genetic model, the fruit fly, Drosophila melanogaster.

We have a remarkable capacity to precisely coordinate body movements. Imagine how effortlessly we reach an arm to grab a cup. Because the cup’s shape/location and our body’s condition are never identical, this seemingly simple motor task requires a sophisticated feedback control; our brain constantly monitors the cup’s and arm’s states and reflects the sensory feedback on generating proper motor commands. Sensory feedback is fundamental, but it has a non-negligible delay that may be problematic for seamless motor coordination.

The theory proposes that the brain predicts the movement outcome in advance to coordinate movement smoothly. Motor deficits in cerebellar patients have suggested the critical role of this predictive capacity. Therefore, our brain likely combines feedback and feedforward (predictive) processing for daily life motor control. Nevertheless, even simple motor tasks recruit many brain areas, complicating our effort to link neural activity and body movements causally.

Fine motor control is not a privilege for humans. It is necessary for many species to survive in an unpredictable environment. We might overlook how elegantly tiny insects control their motor apparatuses (Video 1). Recent advances in optical sensors and computar vision enable us to carefully track and characterize their "subtle" (from our perspective) body movements (Video 2).     

The fruit fly, Drosophila, is a highly tractable genetic model where we can repeatedly record specific neurons' activity across individuals. Moreover, we can selectively manipulate cellular and molecular activities within the genetically identified neural population. The recent open-source whole-brain connectomics has made Drosophila further attractive to systematically dissect their anatomical circuits at a single-cell resolution (Video 3). 

I have been investigating how the tiny Drosophila brain monitors and controls body movements by combining these unique tools with in vivo electrophysiology and optical imaging.

My first postdoc project revealed that a class of visual neurons processing self-generated visual flow is sensitive to self-walking movements even in darkness [Fujiwara et al., 2017] (Video 4). These neurons are selective to the direction of the fly's turn as they are selective to the direction of visual flow; for example, the right-side population of these neurons responds upon a left turn and to a rightward visual flow produced by the left turn (Fig. 1). These motor-related and visual signals are congruently integrated to faithfully estimate self-movement. Interestingly, genetically activating the right-side neurons promotes rightward steering. Thus, the compact visual-motor circuit converts the reliable estimation about ongoing course deviation into a corrective motor command to maintain the straight walking course. 

My second postdoc study revealed that the same class of visual neurons monitors each stepping movement during walking (apart from the above-mentioned turn detection) [Fujiwara et al., 2022] (Fig. 2). This step modulation is amplified upon a step with a larger course deviation. This short (just a few hundreds of milliseconds!) neural activity change is correlated with how much the fly compensates for the deviation on the next step. These observations suggest that visual neurons' activity, or the fly's brain activity in general, contributes to step-by-step rapid course control. In the same paper, I propose that this step-based neural modulation also underlies walking speed representation over several steps. These findings open the possibility to use Drosophila to examine the neural mechanisms for high-end motor control at a cellular and sub-second resolution.

My lab at Riken will continue to challenge big questions in motor control using Drosophila by combining electrophysiology, two-photon calcium imaging, behavior, genetics, engineering, and quantitative analysis. Revealing the detailed mechanisms in this simpler model may provide ground-breaking insights into the universal principles across species, including humans, thereby contributing to therapy for patients.

If you are interested in our lab's research, please JOIN US!! 

Video 1. Flies can do better than you think

Video 2. Pose estimation of tiny flies (Ref.)

Video 3. Fly brain connectomics (Ref.)

Video 4. Neural activity in a walking fly

EPhysSpontWalkDark-100Hz-33-46 (3)_mpeg4.mp4

Fig. 1 Walking course control system

Fig.2 Stepping neural modulation