CROSS-CORRELATION ANALYSIS REVEALS LAMINAR DIFFERENCES IN THALAMOCORTICAL INTERACTIONS IN THE SOMATOSENSORY SYSTEM

1. Spontaneous and stimulus-induced activity were recorded from corres ponding somatotopic representations in the ventroposterolateral nucleu s (VPL) of the thalamus and primary somatosensory (SI) cortex of intac t, halothane-anesthetized cats. Thalamic and cortical neurons with ove rlapping receptive fields on the hairy skin of the forelimb were excit ed by a series of interleaved air jets aimed at multiple skin sites. 2 . The laminar locations of 68% (240 of 355) of the neurons recorded in SI cortex were histologically reconstructed and responses of these 24 0 SI neurons were analyzed with respect to responses recorded from 118 thalamic neurons. Maximum responsiveness during the initial onset (1s t 100 ms) of air jet stimulation was similar for neurons distributed t hroughout all layers of SI cortex (2-4 spikes per stimulus) and did no t differ significantly from VPL responses. During the subsequent plate au phase of the stimulus, VPL neurons discharged at a mean rate of 19. 0 spikes/s and neurons in cortical layers II, IIIa, mb, and IV dischar ged at similar rates. Mean responsiveness during the plateau phase of the stimulus was significantly reduced among neurons in cortical layer s V and VI and only averaged 7.1 and 3.9 spikes/s, respectively. 3. Re sponses recorded simultaneously from pairs of thalamic and cortical ne urons were analyzed with cross-correlation analysis to determine diffe rences in the incidence and strength of neuronal interactions as a fun ction of cortical layer. Among 421 thalamocortical neuron pairs displa ying stimulus-induced responses, 68 neuron pairs exhibited significant interactions during air jet stimulation. A laminar analysis revealed that 28% (45 of 163) of the neurons in the middle cortical layers disp layed significant interactions with thalamic neurons, whereas only 14% (13 of 92) of superficial layer neurons and 6% (10 of 166) of deep la yer neurons were synchronized with thalamic activity during air jet st imulation. When thalamocortical efficacy for different layers of corte x was plotted as a cumulative frequency distribution, the strongest in teractions in the middle cortical layers were twice as strong as inter actions involving the superficial or deep cortical layers. 4. More tha n 70% of stimulus-induced interactions involved thalamic discharges fo llowed by subsequent cortical discharges and the majority of these int eractions involved interspike intervals of less than or equal to 3 ms. Nearly 75% (27 of 37) of interactions in the thalamocortical directio n that involved cortical neurons in layers IIIb and IV transpired with in a 3-ms interspike interval. For interactions with superficial or de ep cortical layers, the proportion of thalamocortical interactions tra nspiring within 3 ms was only 58% (7 of 12) and 33% (2 of 6), respecti vely. 5. Cross-correlation analysis of spontaneous activity indicated that 124 pairs of thalamic and cortical neurons displayed synchronous activity in the absence of sensory stimulation. A laminar analysis ind icated that similar proportions of cortical neurons in each layer were synchronized with thalamic activity in the absence of cutaneous stimu lation. Thus 27% (44 of 163) of middle layer neurons, 30% (28 of 92) o f superficial layer neurons, and 31% (51 of 166) of deep layer neurons displayed spontaneous interactions with thalamic neurons. The tempora l pattern of spontaneous activity was examined with autocorrelation an alysis to determine whether neuronal oscillations were essential for c oordinating thalamic and cortical activity in the absence of periphera l stimulation. Only 18.5% (23 of 124) of spontaneous interactions betw een thalamic and cortical neurons were associated with periodic activi ty, which suggests that thalamocortical synchronization occurs before the constituent neurons begin to oscillate. 6. The influence of sensor y stimulation on spontaneous interactions was examined in 31 pairs of thalamic and cortical neurons that exhibited interactions during prest imulus and stimulus intervals. A matched-sample analysis indicated tha t thalamocortical connection strength was significantly greater during cutaneous stimulation than during spontaneous activity. Further analy sis of peak half-widths in the cross-correlograms also showed that cut aneous stimulation produced significant decreases in the temporal vari ability of thalamocortical interactions. 7. These findings indicate th at neurons in VPL have stronger functional connections with SI neurons in layers IIIb and IV than with neurons in other cortical layers. The relative incidence of interactions with neurons in superficial and de ep cortical layers, together with laminar differences in the rate of s timulus-induced activity, suggests the presence of serial projections from middle to superficial layer neurons. These results are consistent with anatomic evidence, and the relationship between thalamocortical physiology and anatomy is discussed further in the text.