(a) Histogram of waveform parameters from single units recorded in freely behaving mice show a bimodal distribution of peak-to-trough time across subcortical units (n = 28 units, 5 mice). Units with peak-to-trough times under 200 μs were categorized as Narrow (putative TRN), and units over 200 μs were categorized as Wide (putative thalamocortical [TC]). (b) Putative thalamocortical (Wide) units consistently decrease their firing rates during laser stimulation. Mean firing rate in 500 ms bins, shaded region is std. err. across units. (c–e) Heterogeneous firing rates in TRN during stimulation: 4 units strongly increase their firing rates, whereas 10 units decrease their firing rates. The modulation in firing rate is strongly time-locked to laser onset and offset. Shaded regions are std. err. across units. (f) Phase-locking effects across all subcortical units show that most become phase-locked to the slow waves during TRN stimulation. Circles mark the change in phase-locking for each unit; error bars show median change with 25th and 75th quartiles. (g) The phase distribution of putative TRN neurons is broad, with different neurons exhibiting different preferred phases. (h) Peak phase-locking values of putative TC neurons show a tight distribution (Kurtosis = 3.99, n = 11 units), indicating that nearly all putative TC neurons show similar phase-locking to the LFP. Putative TC phase-locking is more consistent across units than putative TRN phase-locking (in b; Kurtosis = 0.24, n = 17 units, group difference = 3.74, significant at α = 0.05 from bootstrap resampling). (i) Example spike rasters around laser onset from 3 single units. Units were not recorded simultaneously; each raster is an independent example. (j–l) Example ISI histograms in single units. (m) Example of parameter estimates from generalized linear model for one unit shows the contribution of recent (<10 ms) spike history increases during stimulation. Shaded regions are std. err.
DOI:
http://dx.doi.org/10.7554/eLife.08760.014