Fig. 3
Working memory deficits are associated with impaired 4 Hz VTA-PFC synchrony in D2R-OE mice. a Example of simultaneously recorded LFP traces in the VTA (top, green) and PFC (middle, black). Thin lines represent the raw LFP whereas thick lines represent the LFP signal filtered between 3 and 6 Hz. The bottom graph shows the normalized 3–6 Hz filtered signals in the two structures, revealing coherent 4 Hz oscillations. Scale bar: 250 ms; 0.2 mV. b, c LFP coherence in sample and choice phases in control (b) and D2R-OE (c) mice. d Four Hertz coherence (calculated by averaging between 3 and 6 Hz) during sample and choice phases in the two genotypes. A selective increase in 4 Hz coherence was observed in control (n = 19, ***p < .001, sign-rank test) but not D2R-OE (n = 17, p = 0.62) animals. e Change in 4 Hz VTA-PFC coherence (choice minus sample, Δ) plotted against learning rate (days to criterion) for each D2R-OE (top) and control (bottom) mouse. Animals that showed a greater increase in coherence in the choice phase learned the task more rapidly (D2R-OE: r = −0.81, p < .0001; Control: r = −0.63, p < .01, Pearson’s correlation). f Change in coherence from sample to choice (top) and behavioral performance (bottom) during the first 40 trials of training (4 days). Control animals showed an increase in coherence only during trials 11–40 (n = 19, ***p < .001, **p < .01, sign-rank test). g Change in coherence on first day of training and the day when criterion performance was reached. D2R-OE mice did not show an increase in coherence during training, even on the day when criterion was reached (n = 17, p = 0.62). Shaded regions and error bars represent mean ± s.e.m measured across animals
