DELAY-TUNED NEURONS IN THE MIDBRAIN OF THE BIG BROWN BAT

1. The auditory midbrain in Eptesicus contains delay-tuned neurons tha t encode target range. Most delay-tuned neurons respond poorly to tone s or individual frequency-modulated (FM) sweeps and require combinatio ns of FM sweeps. They are combination sensitive and delay tuned. The i ndex of facilitation (IF), a coefficient measuring combination sensiti vity for individual delay-tuned neurons, ranged from 0.14 to 1.0, with an average of 0.64 +/- 0.24 (mean +/- SD). Of the 33 facilitated resp onses from 29 neurons, 23 (70%) exhibited IFs > 0.5, which corresponds to a facilitated response 3 times greater than the sum of the respons es to individual pulse and echoes. Thus the responses of midbrain dela y-tuned neurons are highly combination sensitive. 2. The response of m idbrain delay-tuned neurons is phasic, with an average of 0.7 +/- 0.4 action potentials elicited per optimal pulse-echo pair. Thus midbrain delay-tuned neurons in Eptesicus act as probability encoders. 3. The d istribution of best echo delays (BDs) of midbrain delay-tuned neurons ranged from 8 to 30 ms. As an ensemble, midbrain delay-rutted neurons encode target ranges of 138-516 cm. There is a basic correspondence be tween the physiologically determined span of midbrain BDs between 8 an d 30 ms and the behaviorally determined borders of the approach (8- to 17-ms echo delay) and search stages (17- to 30-ms echo delay) of the insect pursuit sequence. Midbrain delay-tuned neurons can be separated into two subpopulations on the basis of the difference in distributio ns of the echo best amplitude (EBA) tuning al ED. The BDs of one subpo pulation correspond to the span of search stage echo delays, and the B Ds of the other subpopulation correspond to the span of approach stage echo delays. 4. EBAs of neurons in each subpopulation are tailored to the specific perceptual requirements of the corresponding behavioral stage. EBAs of midbrain neurons tuned to echo delays between Ij and 30 ms (N = 12) correspond to the search stage and are suited to the requ irements of target detection. EBAs of midbrain neurons tuned to echo d elays between 17 and 30 ms (N = 21) correspond to the approach stage a nd are suited to the requirements of target size discrimination. 5. Th e best FM sweeps for the pulse (PFM) and echo (EFM) were determined fo r each midbrain neuron. PFMs appear to cluster at frequencies correspo nding to the three harmonic peaks in the emitted pulse power spectra. Furthermore, pulse best amplitudes (PBAs) determined at these PFMs cor respond to the amplitude of an emitted pulse assuming middle ear atten uation but without neural attenuation. These response properties refle ct properties of an emitted pulse. 6. In contrast, EFMs are distribute d continuously over a range from 30 to 80 kHz. Furthermore, there is a correlation between EFMs with the lowest EBAs and the frequencies ass ociated with the lowest relative power in the emitted pulse power spec tra. This result suggests that the frequency analysis of pulses and ec hoes by midbrain delay-tuned neurons is different and that delay-tuned midbrain neurons may encode spectral notches in returning echoes caus ed by reflections from multiple targets at different ranges. 7. The di fference between the PEA and EBA versus the difference between the pul se minimum threshold (PMT) and the echo upper threshold (EUT) can be u sed to determine the degrees of overlap between the pulse and echo amp litude tuning of each neuron. The extremes of pulse-echo amplitude tun ing overlap (-27 dB) and lack of pulse-echo overlap (42 dB) suggest th at paradoxical latency shift is not the mechanism underlying delay tun ing in Eptesicus because the PMT should be close to the EUT at BD of t hese midbrain neurons in Eptesicus. In addition, the majority (26 of 3 3, 79%) of EFM:PFM ratios were > 1.5 or < 0.5. This is further evidenc e against paradoxical latency shift as the mechanism underlying delay tuning in this species because the PFM and EFM of the each neuron shou ld be similar if the delay tuning mechanism was paradoxical latency sh ift.