CRICKET INSPIRED PERCEPTION AND AUTONOMOUS DECISION AUTOMATA

For the integration of in vitro networks of insect neurons on uncoated, silicon oxide extracellular recording sites we required a cell culture protocol that allows a low density, long-term stable neuronal cell culture of Gryllus bimaculatus and Locusta migratoria. Retaining of neurites and electrophysiological activity were required to assure the functionality of reconstructed neural networks. A high quality cell culture protocol adapted from methods described in literature was established for cricket (Gryllus bimaculatus) as well as for locust neurons (Locusta migratoria).

Sensory ecology has recently emerged as a new focus of the study of how organisms acquire and respond to information from and about their environment. Many sensory scientists now routinely explore the physiological basis of sensing such as vision, chemoreception or echolocation, in an ecological context. Understanding how the animals actually use the information gathered by their sensors is an essential prerequisite to successful and powerful bionic implementations. This needs to gather proper quantifiable information on the physical and ecological constraints that influence the way in which organisms optimise their performance of stimulus acquisition and processing. We tackled this issue using the anti-predatory fluid motion sensing system of crickets, through a combination of field and laboratory experiments at different levels of the biological chain. The major results in this part of the project are: - The characterisation and quantification of the predation pressure by spiders, the main natural predators of wood crickets. Predation pressure is most important on early stages. Juvenile crickets differ from adults on a behavioural level by hiding under the leaves. - The development of a piston device that faithfully mimics predatory strike showed that spiders adopt two potential tactics (attacking fast or very slowly) to counteract the air movement detection system of crickets. - The understanding of the relative importance of the airflow around the cercus and the role of interactions between receptors enabled us to understand the design of a cercal fovea in cricket. It is characterised by a high density of non-interacting short hairs located at the base of the cercus where sensitivity to air currents is the highest. - The overall structure of the cercal canopy is tuned to detect predators in a time-frequency-intensity space both as far as possible as at close range.

To induce neurite guiding in insect cell cultures cell attractive biomolecules were printed using micro-contact printing (µCP) onto a cell repellent background in order to create a high contrast between the two areas. In cooperation with the Deutsche Wollforschungsinstitut at the RWTH Aachen we developed a new procedure to bind concanavaline A (con A) covalently on a background of so-called star PEG (polyethyleneglycol). PEG is known to hinder protein binding and therefore it should reduce cell adhesion. As a result cricket neurons adhered only on con A areas and neurite outgrowth was restricted to areas modified with con A.

As in most biological sensors, the cricket hairs sensors investigated in the IST-2001-34718 CICADA project show a high level of integration between the "hardware" and the "software" sides. Hair-based mechano-sensors in insects and arthropods achieve very high sensitivities, being capable of detecting acoustic signals at thermal noise levels. Typical threshold levels for neural response are associated with angular displacements of the hairs of the order of 0.001 degrees response to stimuli with air velocities of the order of mm/s or less. To a significant extent this high sensitivity is due to the detailed morphology of the socket-hair subsystem and to the mechanical properties of the tissues involved. The organization of the tissues making up the sensing unit (hair + socket + hair-base supporting membrane) and their mechanical properties can provide local amplification and filtering of the stimulus. The hair and the socket are stiff composites made of chitin fibres bonded together by protein and phenolic compounds. The orientation of the fibres in the hairs is longitudinal but the orientation of the fibres in the socket is yet to be determined. In the regions of the exoskeleton between hairs the chitin fibres are organized plywood fashion. The membrane, which anchors the base of the hair to the socket, is likely to by made of resilin, a rubber-like protein found also in the flight organs of insects. The elastic modulus of this membrane determines the stiffness of the rotational spring supporting the hair and, in turn, this is critical for the mechanical response of individual hairs to steady state or oscillatory airflow stimuli. The objective of WP2 was to characterize as fully as possible the relevant morphological aspects and mechanical properties of tissues, which make up the mechanosensors and to investigate the stimulus-response characteristics of single biological sensors. A data set of correlations between various important morphological and geometrical parameters of the cricket's hair sensing organs has been established using a variety of microscopy techniques. The length of hairs, the easiest geometrical feature to measure using these methods, has been cross-correlated to the diameter at the base, the degree of taper along the length, the diameter of the socket and the ellipticity of the socket base. This last parameter is relevant to the polarity of the hair response, i.e. to the fact the sensing hairs deflect preferentially in one particular plane, with respect to axis of the cerci, determined by the geometry of the elliptical insertion of the hair at the base of the socket. Also, the orientation of this preferential plane with respect to the position of two (occasionally three) campaniform organs associated with at the base of the socket has been verified. The characterization of the elastic properties of the hair-supporting membrane has been obtained using vibration methods and high-speed camera filming. As far as the response of single biological sensors to stimuli is concerned, the combined use of Laser Doppler Vibrometry and high speed camera has shown that over a wide range of input frequencies of the oscillatory stimulus (from a few Hz to several kHz), the hairs 'follow' the stimulus and appear to be in phase with it.

Despite the advancement of mankind in many areas of technology it is still challenging for engineered systems to compete with biological systems. For example: the auditory capabilities of bats to perceive their environment, locate prey and navigate at high velocities through complex surrounding (e.g. leafed brushes and trees) has no manmade equivalent. Likewise the sensitivity of hair-based auditory mechano-sensors, found on e.g. crickets, to detect acoustic signals at thermal noise levels is astounding. It is this kind of performance that raises interest in biological systems and forms an invitation to critically reassess engineered systems with the purpose to see how biology can form an inspiration for manmade systems: e.g. biomimetics. Within the IST-2001-34718 CICADA project workpackage 4 was concerned with the fabrication of flow-sensing electro-mechanical elements by silicon micro-machining technology (MEMS) and the integration of many of these sensors into arrays. Here the challenges were specifically in a) choosing and implementing proper sensing principles, b) assembling a proper processing sequence that allows for integration of long hair-like structures with thin minute sensor structures and c) optimising all parts of the sensors and sensor-arrays to achieve sufficient sensitivity. In designing and fabrication the input from other work packages is used to form optimum topologies. In the CICADA project we have successfully fabricated artificial hair-sensors and hair-sensor arrays. These sensors are based on drag-force mediated rotations of membranes resolved by capacitive measurements. The devices were the first in literature reported sensors of this kind to show acoustic sensitivities. Although a comparison of the artificial sensors with those of crickets is delicate we approximate that our reported sensors (with hair-length of 470µm) are a factor of 10^4-10^5 less sensitive than those of crickets. However, improvements in mechanical design (lower suspension stiffness), improved electronics and the use of the longer hairs we expect to be able to improve the sensitivity by a factor of 10^3-10^4 in the future. The technology developed in this project allows for future developments addressing: 1) spatio-temporal flow and sound measurements, 2) adaptive control of sensitivities, 3) extension to sensors operating in fluids, 4) use of mechanical filtering (e.g. like in the mammalian cochlea) and 5) beneficial use of noise (stochastic resonance). We are happy that the EU has shown a willingness to fund this type of research in the successor project Cilia (Customised Intelligent Lifelike Arrays) in which we will address some of the points listed above.

Rezultaty

Udostępnij tę stronę

Pobierz