Cross-correlation analysis of cuneothalamic interactions in the rat somatosensory system: Influence of receptive field topography and comparisons with thalamocortical interactions

1. We simultaneously recorded neuronal responses to cutaneous stimulation from matched somatotopic representations in the nucleus cuneatus and ventrobasal complex of intact, halothane-anesthetized rats. A total of 95 cuneate and 86 thalamic neurons representing hairy skin on the forelimb were activated by hair movements produced by air jets at multiple skin sites. Mean responsiveness was higher among neurons in nucleus cuneatus (34.4 spikes per stimulus) than in thalamus (23.7 spikes per stimulus), a result that was consistent with the greater proportion of 'sustained' responses recorded in nucleus cuneatus (80%) than in the thalamus (62%). 2. Cross-correlation analysis of 166 pairs of cuneate and thalamic neurons showed that 56 neuron pairs displayed time-locked correlations in activity that were characterized primarily by excitatory interactions (44 pairs) or a combination of excitatory and inhibitory interactions (10 pairs). Unilateral interactions in the cuneothalamic direction (31 pairs) and reverse direction (11 pairs) were observed, as well as multiphasic interactions in both directions (14 pairs). Most excitatory interactions involved intervals of 1-7 ms between successive cuneate and thalamic discharges, whereas most inhibitory influences involved intervals >7 ms. Connection strength, defined by the ratio of time-linked interactions to the number of cuneate discharges, varied widely among neuron pairs but was largest for interactions involving interspike intervals of ≤15 ms. 3. The relationship between connection strength and receptive field topography was analyzed in 103 cuneate-thalamic neuron pairs. The region of skin shared by both neurons varied substantially among neuron pairs and the probability of detecting interactions increased proportionately with larger amounts of receptive field overlap. Neuron pairs with moderate (25-50%) amounts of receptive field overlap had connection strengths 3-4 times greater than neuron pairs with minimal (0-25%) overlap. Connection strength was essentially identical, however, for neuron pairs with moderate or large (>50%) amounts of overlap. 4. Cuneate-thalamic neuron pairs displaying functional connections were usually tested at multiple peripheral sites, but only 37% (18 of 49) of these neuron pairs displayed interactions at more than one stimulation site. Stimulation at different sites altered the timing of interactions in seven neuron pairs, including three that showed timing shifts across time zero in the cross-correlation histogram. In neuron pairs displaying interactions at multiple sites, connection strengths for 67% of the cases were strongest when stimulation was delivered within the region of receptive field overlap. Among 31 neuron pairs displaying interactions at a single stimulation site, 57% of these sites were in the central region of overlap. 5. In 10 experiments we tested thalamic responsiveness before and after making a discrete electrolytic microlesion at the cuneate recording site. In most cases thalamic neurons became less responsive and the decrease in activity was proportional to the maximum connection strength associated with each pair of recording sites. 6. We compared cuneate-thalamic interactions with thalamic-cortical interactions obtained from a previous study using identical protocols. Stimulus-induced responses were largest in nucleus cuneatus and were progressively smaller in the somatosensory thalamus and cortex. Analysis of the temporal structure of the cross-correlation histograms revealed that cuneothalamic interactions had less variability and involved shorter temporal intervals than thalamocortical interactions. These differences were consistent with evidence suggesting that cuneate-thalamic interactions are more likely to involve monosynaptic connections. 7. Cuneate- thalamic and thalamic-cortical interactions were analyzed during spontaneous and stimulus-induced activity. Interactions between thalamic and cortical neurons were equally likely during spontaneous or stimulus-induced activity, whereas cuneate-thalamic interactions occurred almost exclusively during stimulus-induced activity. These results showed that cuneothalamic and thalamocortical projections are differentially involved in relaying sensory or state-dependent information. Analysis of stimulus-induced interactions suggests that cuneothalamic synaptic connections are more efficacious than thalamocortical synapses. These results are consonant with known anatomic differences between cuneothalamic and thalamocortical connections and are discussed in the text.