Faculty Bibliography

Neural circuits often face the problem of classifying stimuli into discrete classes and making decisions based on such classifications. Neurons of such circuits can be, quite generically, separated into three groups: sensory cells, whose activity is highly correlated with the stimulus; decision-making neurons, which are responsible for classifying the stimulus; and motor neurons, which implement the results of such classifications. Although the former and the latter groups can be easily found experimentally, by e.g. studying correlations with stimulus and response, there is no acceptable experimental procedure of identifying the decision-making neurons. In this work we consider three types of problems of finding decision-makers experimentally. In the problems of Type I the responses of neurons to a given set of stimuli as well as the topology of underlying network is known. In Type II problems, the responses of neurons are known, but the neuronal circuit is unknown. In the problems of Type III neither the neuronal responses nor the underlying circuit is known, but the researcher is permitted to knock out appropriate parts of circuit and explore changes in the motor response. In this work we put forward a set of experimental strategies, which allow to find parts of neuronal circuits, whose primary goal is making perceptual decisions. The solutions of the three types of problems are tested on surrogate neuronal circuits, in which the identities of decision-makers are known

A novel system for the generation and measurement of a two dimensional wind stimulus is proposed and described. This system was used to investigate the wind sensation of the American cockroach. The new aspects of this system are (a) a pair of computer driven wind tunnels that are shown to produce non-turbulent flows and (b) a novel fiber optic wind detector that measures both amplitude and direction of the wind. Winds can be produced and measured in behaviorally relevant frequency and amplitude ranges without perturbing the airflow. The combination of both the wind generation system and wind detector makes the system very flexible and allows the generation of stimuli with any given spectrum. The two dimensional wind stimulus is shown to be very reproducible. The wind detector is independent of the wind generation system so it can be used for measuring natural winds as well. Experimental data obtained on the cockroach are presented.

We report experimental studies of parametric excitation of second sound (SS) by first sound (FS) in superfluid helium in a resonance cavity. The results on several topics in this system are presented: (i) The linear properties of the instability, namely, the threshold, its temperature and geometrical dependencies, and the spectra of SS just above the onset were measured. They were found to be in good quantitative agreement with the theory. (ii) It was shown that the mechanism of SS amplitude saturation is due to the nonlinear attenuation of SS via three wave interactions between the SS waves. Strong low-frequency amplitude fluctuations of SS above the threshold were observed. The spectra of these fluctuations had a universal shape with exponentially decaying tails. Furthermore, the spectral width grew continuously with the FS amplitude. The role of three and four wave interactions are discussed with respect to the nonlinear SS behavior. The first evidence of Gaussian statistics of the wave amplitudes for the parametrically generated wave ensemble was obtained. (iii) The experiments on simultaneous pumping of the FS and independent SS waves revealed several effects. Below, the instability threshold, the SS phase conjugation as a result of three wave interactions between the FS and SS waves was observed. Above the threshold two interesting effects were found: a giant amplification of the SS wave intensity and strong resonance oscillations of the SS wave amplitude as a function of the FS amplitude. Qualitative explanations of these effects are suggested

Superconducting Pb-Au thin film bolometers for second sound (SS) measurements near the lambda -point have been made on 8 mum diameter glass fibers. The small diameter of the fibers provides negligible cross-section to SS waves as close as similar to 10(-5) K to the lambda -point (for f = 10 kHz), thus eliminating reflections that may unduly complicate the wave distribution in the experimental cell. The bolometers have isotropic sensitivity to SS waves in the plane perpendicular to the fiber. The operating temperature of the bolometers can be adjusted by changing an external magnetic field and/or by varying the bias current, while still maintaining a figure of merit of similar to 30. A SS emitter producing planar waves was made using the same fibers. (C) 2001 Published by Elsevier Science Ltd

This work describes an experimental system aimed at addressing the problem of how multiple cells code features of sensory information. These experiments on the escape response of the American cockroach, Periplaneta americana, enable a quantitative approach to the problem of collective coding due to good control of the stimulus and simultaneous recordings from a large fraction of the coding cells. The following results are presented: (a) some cells code for features of the wind that are not simply directional in nature, (b) compound events, defined as simultaneous firing of two cells, represent features of the stimulus that are different than each of the individual cells. (C) 2001 Published by Elsevier Science B.V

Neurogenesis in central olfactory processing centers is a universal design feature of olfactory systems. The procerebral (PC) lobe of the Limax cerebral ganglion, a likely site of odor memory formation, exhibits substantial neurogenesis after hatching. The PC lobe has oscillatory dynamics of its local field potential and propagates activity waves originating at the apical pole. In Helix substantial neurogenesis occurs in the PC lobe after hatching, accounting for as much as 80% of the cell population in the PC. We ask how odor inputs and ultimately odor learning affect neurogenesis and synaptic connectivity of new neurons in the Limax PC lobe. We soak newly hatched Limax in bromodeoxyuridine (BrdU) for 90 min and assess incorporation into DNA of new neurons using immunocytochemical labeling of BrdU in single-stranded DNA. The site of neurogenesis in the PC lobe is a band-like region spanning the full width of the lobe at the very apex. The band of BrdU labeled neurons is found more basal to the apex as lobes are fixed at later and later times after labeling as the apex is continually growing by asymmetric cell divisions. Only cells on the apical side of the band of neurogenesis exhibit BrdU labeling. Unilateral nose removal for 1 month results in a smaller PC lobe and an apparent reduction in neurogenesis. The neuronal bands resulting from synchronous neurogenesis will be compared with odor memory bands labeled by learning-dependent uptake of Lucifer yellow (LY). It is striking that the wavefront of the propagating activity band is parallel to the long axes of both the band of neurogenesis and the LY-labeled odor memory bands

A system for presentation of olfactory stimuli for studying the learning dynamics in Limax procerebral (PC) lobe should combine the following features: i) presentation of multiple odorants, ii) reproducibility of odorant presentation for many hours, iii) minimal damage to the olfactory organ, iv) accurate measurement of the timing of stimulus presentation, v) measurement of odorant concentration. We built a new system that satisfies all the above conditions. The aqueous odorant presentation system was combined with electro-olfactogram measurement. Eight 100 mum ID capillaries lodged inside a large suction capillary supplied one saline and seven saline-odorant solutions to the inside of its tip. Part of the olfactory epithelium is sealed into the tip of the large suction capillary. The aqueous odorant stimulus is presented by simultaneously switching off the saline line and switching on one of the seven saline-odorant lines. An electrode positioned inside the large suction capillary monitors the currents generated by olfactory receptors, thereby giving the onset time for information processing in the rest of olfactory system. Combining this technique with a newly developed protocol for in-vitro training of the isolated nose-brain preparation will yield new insights into the learning dynamics of the Limax PC lobe

We present the first experimental observation of phase conjugate (PC) second sound (SS) waves superfluid helium. The main feature of a PC wave is that its phase is complex conjugate to the Incident wave phase. It is generated and amplified as a result of nonlinear interaction between an incident SS wave and a first sound (FS) pumping wave. At FS amplitudes larger than the threshold value, a parametric instability, i.e., spontaneous decay of a FS wave into two SS waves, takes place. Three main, theoretically predicted, features of phase conjugate waves were verified experimentally