Figure 2. Spindle-burst oscillations require retinal and thalamic activity.
(A) Blockade of spontaneous retinal waves reduces thalamic firing and spindle-burst oscillations. (A1) Ocular injection of glutamate receptor antagonists APV and CNQX to silence retinal activity with simultaneous recording in LGN. (A2) Representative LGN MUA (>300 Hz) before (control) and after retinal silencing. (A3) Percent change in multi-unit spike rate and continuity of activity during retinal blockade (n = 6, Wilcoxon signed-rank test for difference from pre-silence, p=0.0313, p=0.0313). Spike continuity is calculated as the proportion of periods containing at least two spikes with interval of less than 500 ms. (A4) Analysis of event duration. Cumulative distribution is shown in solid lines (left y-axis). 95% confidence interval (CI) is shown as shaded area (n = 856 and n = 52, two sample KS test, p=0.0193). Proportion distribution is additionally plotted as a bar graph to aid visualization of event duration (right y-axis). (A5) Population mean of normalized spike spectra. Silencing retina reduced spindle-burst frequencies (n = 6, permutation test, p<0.05 between 23.9–25.5 Hz, bar = frequencies significantly different by permutation test). (B) Silencing retinal activity reduces cortical firing and spindle-burst oscillations. (B1–5) As for A1–5 but for L4 of VC (B2: VC LFP (1–150 Hz); B3: n = 6, Wilcoxon signed-rank test, p=0.0313, p=0.0313; B4: n = 1109 and 189, two sample KS test, p=10–12; B5: n = 6, permutation test, p<0.05 between 14.4–54.6 Hz). (C) Silencing thalamic activity with muscimol while recording in VC. (C1–5) As for B1–5. Larger reduction in spontaneous firing, event duration and spectral power than for retinal silencing suggests residual activity in thalamocortical loop after retinal silencing can generate small amounts of spindle-burst activity. (C2: VC LFP (1–150 Hz); C3: n = 6, Wilcoxon signed-rank test, p=0.0313, p=0.0313; C4: n = 811 and 377, two sample KS test, p=10–22; C5: n = 6, permutation test, p<0.05 between 8.6–42.5 Hz). All error bars are SEM unless noted.