The challenge of perceiving the visual world is an ancient design problem. There are three distinct functional requirements for any effective system: animals must detect objects in a large area of the environment, they must achieve sufficient resolution to identify them, and they must reduce the visual information to a level that matches the capacity of their brain to handle it.

    Familiar vertebrate eyes like our own are but one solution. Remarkably, at least 10 distinct designs have evolved. These can be broadly grouped into two classes, each accommodating the competing tasks of high-resolution vision and motion detection. Some animals have large eyes that address both functional requirements in a single structure; this obliges them to allocate a relatively large proportion of their neural resources to the massive processing load that such a system generates. The alternative is to adopt a design that segregates the tasks of motion and form processing into separate, dedicated, systems. The resulting output makes much more modest computational demands, suitable for animals with small brains.

    The seemingly impossible feat of achieving high acuity and large field of view with only modest computational power has been achieved in only one group. Jumping spiders have evolved a unique ‘modular’ design. Six secondary eyes function as motion detectors, functionally analogous to peripheral vision in vertebrates. These combine to produce a field of view of almost 360 degrees. When something interesting is detected, the spider spins around to bring its two large forward-facing primary eyes (analogous to a fovea) to bear. These animals have a body length of just 12 mm and brain only half the size of a honeybee’s, yet their sophisticated visual system attains a visual acuity that is almost an order of magnitude superior to their closest insect rivals and approaches that of primates. This is a stunning demonstration of evolutionary design and miniaturisation that, were it understood, would make our best robotics engineers weep.

    Results will be of considerable theoretical interest because they will permit comparison with other systems that have evolved completely independently. In addition, we aim to make a contribution to the burgeoning field of biomimetic optics. The performance characteristics we will document (high-resolution, wide field of view, and manageable processing power) are precisely those required for a successful artificial visual system.

 

We will combine behavioural and neurophysiological studies to develop a new model system rich as a source of ideas concerning visual sensory processes and their behavioural ramifications. This project will be the first of its kind to directly address motion analysis in jumping spiders and will provide fundamental insights into how small-brained animals solve complex perceptual tasks.