[back]

Neuromechanical Design and Active Sensory Systems in Animals
Malcolm Maciver, Joel Burdick

The field of neuroethology has made tremendous progress in understanding the sensory processing that subserve natural behaviors. Much work remains, however, in obtaining an equally detailed and quantitative understanding of how the mechanics of animals subserve natural behaviors, and in particular, how sensory abilities complement an animal’s mechanical control and locomotory needs and characteristics. In addition to its basic science import, these issues have relevance to engineers seeking to emulate some of key advantages of animal neuromechanical design, such as high maneuverability, and high levels of sensory integration for executing behaviors under changing and uncertain conditions. In this work we study how motion and sensing are integrated in the weakly electric fish.

The black ghost knifefish (Apteronotus albifrons, Fig. 1) is a South American gymnotiform weakly electric fish that can communicate and sense its surroundings by use of a weak (1 mV cm-1 near the body, Fig. 2b), self-generated electric field. Nearby objects that differ in electrical conductivity from the surrounding water create localized voltage perturbations that are sensed by ≈14,000 transcutaneous electroreceptor organs scattered over the body surface (Fig. 2a). In addition to their specialized sensory capabilities, gymnotiform knifefish possess a unique multi-directional propulsion system driven by a ventral ribbon fin that runs most of the length of the body (Fig. 1). By generating traveling waves along the ribbon fin and manipulating the pectoral fins, they move forward, backward, and upward, and can rapidly pitch or roll the body. It is a highly active, predatory fish, hunting for small prey at night in geometrically complex habitats such as flooded forests, and hiding during the day.

Figure 1. Thrust vectors around Apteronotus albifrons for two swimming directions and a roll maneuver. FN is force normal to line of ventral ribbon fin; FP is the force parallel to this fin, and FF is the force generated by pectoral fin. (a) Swimming forward (b) Swimming backwards. Note that the body needs to be negatively pitched to go straight back. (c) Swimming up and back with a roll by forcing fluid down on one side of the body, similar to the motion observed during a stereotypical prey strike (see Figure 3).

Using a combination of experimental and computational techniques, we have investigated two key issues relating to this animal. First, how much do mechanical factors influence the animal’s key behaviors? Second, how does the design and function of the animal’s key sensory systems complement its locomotory and mechanical control needs?

Our investigation is based on prior behavioral studies wherein individual black ghost knifefish were videotaped hunting for small prey (Daphnia magna) under infrared light within a light-tight enclosure. A non-rigid wireframe model of the fish was overlayed onto the video, and a 3D reconstruction algorithm provided the position of the surface of the fish and prey to ±0.5 mm, with a time resolution of 16.7 ms.

Figure 2. (a) Electroreceptor layout for A. albifrons. Data for (a) from [18], covering the surface of a 3D model of the fish according to [19], [20]. There are no sensors on the fins of the fish. (b) Electric field around A. albifrons, midsagittal plane. Data set for (b) courtesy of Chris Assad and Brian Rasnow.

Stereotypical behaviors seen during these prey capture maneuvers included rapid reversals and dorsal rolls. We found, using optimal control theoretic computations, that these commonly observed maneuvers are mechanically optimal. That is, the fish’s behavior is dominated by issues of mechanical efficiency. This finding was accomplished by utilizing an idealized ellipsoidal body model, Kirchhoff’s equations, and an optimal control algorithm for generating synthetic trajectories. Moreover, we found that the sensory distribution of the fish’s primary sense organs complements these mechanically optimal movements.

We have also found that the volume of space that A. albifrons is able to reach within the sensorimotor latency of the animal (time from sensory input to motor output, 120 ms) plus the time it takes to stop the body when going forward at typical searching velocity (115 ms) is nearly congruent with the volume of space within which it can detect the small prey that are its food source (Fig. 6). Remarkably, despite the high variation in sensor density and electric field strength (Fig. 2), this sensing volume forms a very regular cylinder around the body. The unusually high maneuverability of this animal perfectly complements its omnidirectional sensing capabilities.

 

Figure 3. A typical rolling maneuver to reach a laterally-positioned prey. Wireframe represents position of fish as it strikes a prey, from 3D motion capture data. The top snapshot (t=0 ms) is at the time of prey detection, and time increases up to the last snapshot at the end of the sequence. The heavy line on the fish indicates the dorsal edge, the open circle marks the position of the Daphnia magna, and the dotted line indicates the shortest distance from the Daphnia to the body surface. The inset plot on the left shows the roll angle history and current value (filled circle).

Figure 6. Sensory and small-time locomotor volumes around A. albifrons. (a) The sensory signal isosurface for detecting small prey. All points on the surface result in the same net input to the ≈14,000 sensor receptors covering the surface. (b) The stopping-time locomotory volume (all points reachable within 235 ms of movement of the animal), indicated by purple locomotory isosurface, with the sensory isosurface shown in light red.