Excitatory cells in layer 6 have diverse morphologies but fall into broad categories according to their long distance projections (Katz, 1987; Zhang & Deschenes, 1997). Corticothalamic cells (CT) have apical dendritic tufts in layer 4 and local axon collaterals arborising in layer 4. Corticocortical cells (CC) are more varied, including short pyramids, inverted pyramids and bipolar cells, but lack apical tufts in layer 4 and have horizontally directed axon collaterals restricted to layers 5/6. These differences suggest distinct roles within the network.

We recorded and filled pairs of layer 6 cells in 450 -500µm slices from 5-6 week old male rats and adult male cats at 34°C (Anaesthesia: halothane + 60 mg kg^-1 pentobarbitone I.P. for rats; 70 mg kg^-1 α-chloralose + 6 mg kg^-1 pentobarbitone I.V. for cats. Animals were humanely killed. Further details in Thomson & West (2003).

Three firing types were observed in rat: tonic (8 cells), phasic (20), and phasitonic (phasic to small suprathreshold depolarisations but tonic to larger depolarisations, 10 cells). Firing type and morphology were correlated: of 17 CT-like cells, most were tonic (8) or phasitonic (6). In contrast, none of 21 CC-like cells was tonic and 17 were phasic. 1037 pairs of layer 6 cells in rat and 141 in cat were tested for synaptic connections, yielding 37 and 8 connections respectively. Average EPSP amplitude (at -72 mV) was 0.73 ± 0.43 mV (mean ± S.D., n = 22) and transmission failures ranged from 0 -54 % (12 ± 16 %). Most filled cells had CC-like morphology and phasic firing properties.All connections from CC-like cells exhibited short-term synaptic depression: 2nd EPSP amplitudes were 15 -60 % of the 1st at intervals < 20 ms, recovering to ~80 % at 60 -100 ms. The one connection between CT-like cells recorded had qualitatively different short-term properties, exhibiting clear facilitation (> 150 % at 15 ms interval) and post-tetanic potentiation (EPSPs following 6 spike trains at 3 s intervals were 2.3Î those following single spikes).

These data support previous suggestions that CC cells make depressing synaptic connections whereas CT cells make facilitating connections (Beierlein & Connors, 2002; Tarczy-Hornoch et al. 1999). The phasic firing properties of CC cells and their depressing synaptic connections are best suited to signalling changes in incoming activity (novelty detection), whereas the ability of CT cells to fire tonically and their less phasic outputs may enable sustained input both to layer 4 and to the thalamus.

This work was supported by the MRC