Differences in responses to 70 dB clicks of cerebellar units with simple versus complex spike activity: (i) in medial and lateral ansiform lobes and flocculus; and (ii) before and after conditioning blink conditioned responses with clicks as conditioned stimuli.

Activity was recorded from 554 cerebellar units in eleven conscious cats to determine if responses to 70 dB clicks differed in units with simple and complex spike discharges. Effects of region of recording and behavioral state (with click used as a conditioned stimulus for conditioning) were also assessed. Cells with only simple spikes were distinguished from cells that had the following types of complex spike events: Type I-simple or initial spike followed > 1 ms by multiple spikes with baseline displacement (classical complex spikes), Type II--followed < or = 1 ms by spikes with or without baseline displacement (spikes in the absolute refractory period should arise from a separate site of initiation), and Type III-followed by spikes and displacement too close to the baseline noise to distinguish as Type I or II. Among the groups mean baseline activity was greatest in cells with Type I complex spikes, least in cells with Type III complex spikes, and greater in Type II cells than simple cells. Significant increases in activity within 32 ms of presenting clicks were found in the groups of Type II cells and simple cells. These appear to be the main auditory responsive cells of the cerebellar regions studied. Activity of Type II cells best reflected the temporal properties of the click; responses of simple cells had slower onsets (except in flocculus) and longer durations. Responses to click in Type II and simple cells differed in recordings from: (i) lateral ansiform lobe (lateral crus I and portions of crus II), (ii) medial ansiform lobe (medial crus I), and (iii) flocculus. The largest mean responses above baseline in the first 32 ms after click were found in Type II cells of the lateral ansiform lobe with onsets of 8-16 ms. Magnitudes of response differed before and after conditioning and backward conditioning. In the lateral ansiform lobe, the < 32 ms response to click was greater in Type II than simple cells in each state, but showed a greater increase above baseline after backward conditioning when conditioned responses were not produced than after conditioning. The onset of increased activity to click conditioned stimuli in Type II cells of the lateral ansiform region preceded the onset of the blink conditioned response after conditioning, consisted almost entirely of simple spikes, and reflected an increase in magnitude of response as opposed to an increased number of responsive units. After conditioning, an increased number of units in the flocculus responded to click conditioned stimuli in the 16-24 ms post stimulus period. Of the 16 cells with an onset of increased activity at this time, eight showed only simple spike activity. Seven of the remaining eight cells (all Type II) showed a significant increase in conditioned stimulus-evoked complex spiking above the low (usually < 1/s) baseline level of complex spike discharges. The findings support the conclusions that cerebellar units can respond rapidly enough to acoustic stimuli to play a role in auditory as well as motor processing and that the responses to 70 dB clicks differ among cells with simple and complex spike discharges. The differences are influenced substantially by the region of cerebellar recording and the behavioral state. The findings in cells of the flocculus offer the first evidence that complex as well as simple spike activity can contribute to an increased probability of discharge to click as a conditioned stimulus after conditioning.

Physics-Based Haptic Simulation of Bone Machining.

We present a physics-based training simulator for bone machining. Based on experimental studies, the energy required to remove a unit volume of bone is a constant for every particular bone material. We use this physical principle to obtain the forces required to remove bone material with a milling tool rotating at high speed. The rotating blades of the tool are modeled as a set of small cutting elements. The force of interaction between a cutting element and bone is calculated from the energy required to remove a bone chip with an estimated thickness and known material stiffness. The total force acting on the cutter at a particular instant is obtained by integrating the differential forces over all cutting elements engaged. A voxel representation is used to represent the virtual bone and removed chips for calculating forces of machining. We use voxels that carry bone material properties to represent the volumetric haptic body and to apply underlying physical changes during machining. Experimental results of machining samples of a real bone confirm the force model. A real-time haptic implementation of the method in a dental training simulator is described.