In the presence of the metabotropic glutamate receptor (mGluR) agonists, trans-ACPD (50-200 mM) or DHPG (50-100 mM), thalamocortical neurones of the cat dorsal lateral geniculate nucleus (obtained from humanely killed animals) maintained in vitro exhibit stereotypical, small depolarising potentials or spikelets. In the majority of cases (80 %), these spikelets are isolated rhythmic (~10 Hz) events that represent the electrotonic transmission of single action potentials via gap junctions (Hughes et al. 2002).

In cells exhibiting spikelets, we now report that in a small number of cases (20 %) they appear in rhythmic bursts which ride on a small depolarisation, and which we term compound spikelets. As with isolated spikelets, compound spikelets are blocked by carbenoxolone (100 mM) and usually accompanied by dye coupling (75 %). The interburst frequency of compound spikelets is 4.6 ± 0.8 Hz whilst their intraburst frequency is 66 ± 4 Hz (n = 4). These values correspond closely to those exhibited by a novel form of burst firing (interburst frequency range 3-12 Hz; intraburst frequency = 74 ± 5 Hz; n = 52) that occurs at membrane potentials above -50 mV (high-threshold bursting, HTB) in a small percentage (25 %) of TC neurones in the presence of trans-ACPD or DHPG. Therefore we suggest that compound spikelets are generated by electrotonically transmitted high-threshold bursts. Isolated spikelets and compound spikelets differ in their ability to entrain firing: isolated spikelets entrain single spike activity at >15 Hz, whereas compound spikelets entrain HTB at 3-12 Hz. The latter finding suggests that TC neurones should generate synchronized oscillations at 3-12 Hz, a suggestion confirmed by extracellular multi-unit recordings. Furthermore, dual intra-and extracellular recordings performed in the presence of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), DL-2-amino-5-phosphonovaleric acid (APV) and bicuculline methiodide (BMI) directly demonstrate that (i) these oscillations are not dependent on fast chemical synaptic transmission, (ii) they are driven by HTB and (iii) that their frequency is largely determined by how depolarised the TC neurone population is. In conclusion, our results suggest that in the presence of mGluR agonists, TC neurones generate one of two distinct types of firing pattern: single spike activity or HTB, with single spikes being reflected electrotonically as isolated spikelets and HTB as compound spikelets. The combination of HTB and compound spikelets can generate gap junction driven synchronized oscillations at α (8-12 Hz) and u (3-7 Hz) frequencies that may have relevance to the EEG rhythms that characterize relaxed wakefulness and early sleep stages, respectively (data are quoted as means ± S.E.M.).

This work was supported by The Wellcome Trust (grant 37089-98).

All procedures accord with current UK legislation.