Superior sensation: superior colliculus participation in rat vibrissa system

Marie E Hemelt 0 1 Asaf Keller 0 0 Department of Anatomy & Neurobiology, University of Maryland School of Medicine , Baltimore, MD , USA 1 Program in Neuroscience, University of Maryland School of Medicine , Baltimore, MD , USA Background: The superior colliculus, usually considered a visuomotor structure, is anatomically positioned to perform sensorimotor transformations in other modalities. While there is evidence for its potential participation in sensorimotor loops of the rodent vibrissa system, little is known about its functional role in vibrissa sensation or movement. In anesthetized rats, we characterized extracellularly recorded responses of collicular neurons to different types of vibrissa stimuli. Results: Collicular neurons had large receptive fields (median = 14.5 vibrissae). Single units displayed responses with short latencies (5.6 0.2 msec, median = 5.5) and relatively large magnitudes (1.2 0.1 spikes/stimulus, median = 1.2). Individual neurons could entrain to repetitive vibrissa stimuli delivered at 20 Hz, with little reduction in phase locking, even when response magnitude was decreased. Neurons responded preferentially to vibrissa deflections at particular angles, with 43% of the cells having high ( 5) angular selectivity indices. Conclusion: Results are consistent with a proposed role of the colliculus in somatosensorymediated orienting. These properties, together with the connections of the superior colliculus in sensorimotor loops, are consistent with its involvement in orienting, alerting and attentive functions related to the vibrissa system. - Background The superior colliculus has long been studied as a center for visual sensory and motor responses [1-3], and is involved in orienting attention [4-6] and multimodal processing [7-10]. The functions of the colliculus have been characterized principally in relation to vision and eye movements, whereas its role in other sensorimotor transformations has received less attention. We seek to explore the function of the superior colliculus using the vibrissa system, a vital sensorimotor system for the rat. Although previous studies demonstrated the existence of vibrissa-responsive collicular neurons [9,11-13], there have been no systematic, quantitative analyses of their response properties to controlled stimuli. The goal of this study was to correct this deficiency. The superior colliculus is thought to be part of a sensorimotor loop in the rat vibrissa system, receiving input from trigeminal nuclei and projecting to the vibrissa motoneurons [see [14]]. In addition, the superior colliculus interacts with other structures of the vibrissa-barrel system: it receives input from the barrel cortex and from the vibrissa area of the motor cortex [15-17], and receives inputs from the cerebellum and basal ganglia [see [18]]. It is recipro) ce 800 s (m700 t ne 600 n po 500 m o 400 C e 300 v i t is 200 o P P. cally connected with the zona incerta [19], a structure recently shown to be involved in state dependent gating of vibrissal inputs in the thalamus [20-22]. Because of these anatomical relationships, and because of their role in orienting responses, we hypothesized that collicular neurons respond with short latency and strong directional selectivity to vibrissal stimuli. Results Location of vibrissa responsive units We report data from 51 vibrissa-responsive units recorded from 24 animals. We used histological analyses, stereotaxic coordinates and depth readings from the manipulator to determine the location of the recorded units. Due to the size of the lesions, we did not separate intermediate gray and white layers, but divided the colliculus into superficial, intermediate, and deep layers. Vibrissa responsive units were located in the intermediate layers of the lateral and posterior parts of the colliculus. Stereotaxic coordinates were 2.2 to 2.6 mm lateral and 7.6 to 7.9 mm posterior to Bregma, at a depth of 3.2 to 5.3 below the cortical surface. Clusters of vibrissa responsive neurons were interspersed with clusters responding to manual stimulation of other body parts. Vibrissa-responsive clusters ranged in size from 100 to 300 microns along the axis of penetration; it was not possible to reliably determine the cluster size in the horizontal plane without fine-grained mapping in a single animal. The clustering of vibrissa responses may reflect the clustering in the colliculus of afferents from the trigeminal nuclei [23,24]. Waveform morphology Neurons in several brain regions have been classified according to their action potential waveforms [25,26]. For example, cortical vibrissa-responsive units include regular and fast spiking neurons [27,28]. In extracellular recordings, inhibitory, fast spiking neurons have action potentials with a brief ( 180 sec) negativity followed by a brief positivity ( 400 sec), whereas regular spiking neurons have longer waveform components (> 180 sec and > 400 sec). The superior colliculus contains both excitatory (glutamatergic) and inhibitory (GABAergic) neurons, which may segregate into similar categories [18,29]. To test this, we measured the two components of their extracellularly recorded waveforms, and plotted the positive component against the negative component (Fig. 1). All data points (51 data points, 12 overlapping) formed a single cluster with values encompassing those of cortical regular and fast spiking units. While we cannot exclude the possibility that our sample included a heterogeneous population of neurons, these findings suggest that our recordings could not distinguish neuronal subclasses based on their waveforms. Negative Component (msec) FwNiaegvuuerrfooenram1l smubocrlpahssoelosgcyould not be distinguished by their Neuronal subclasses could not be distinguished by their waveform morphology. The inset shows representative waveforms, with the negative (N) and positive (P) components indicated. When the positive component was plotted against the negative component, data points formed a single cluster. Spontaneous activity Vibrissa-responsive neurons had low or no spontaneous activity (e.g. Fig. 2AC; 44/48 neurons had no spontaneous activity, mean was 0.01 0.006 Hz). Therefore, to avoid biasing our recording to spontaneously active units, we searched for units while manually stimulating the vibrissae. We did find many spontaneously active, albeit non-responsive units near these relatively quiescent, responsive neurons. We found no neurons for which stimulation evoked a decrease in spontaneous activity. Response latency The superior colliculus receives input from a variety of structures that process vibrissa inputs, including the trigeminal nuclear complex and the neocortex (see Background). We reasoned that if collicular responses represent direct inputs from the trigeminal nuclei, these responses would have relatively short latencies, similar to those of thalamic neurons that receive trigeminal inputs. Figures 2AB depict responses recorde (...truncated)