(A) Example coronal sections showing TH-IR in the dorsal striatum of unilaterally depleted mice. Scale bar, 100 µm. (B) Quantification of the percent TH remaining (relative to control) in the ipsilateral (‘Ipsi’) and contralateral (‘Contra’) hemispheres of unilaterally depleted mice. (C) Box plots of firing rates from all neurons recorded in the ipsi and contra hemispheres. Grey squares indicate animal medians. KW, χ2(3) = 5.488, p < 0.0001, pairwise, **p < from Ctl. (D) Box plots of CVISI from all neurons recorded in the ipsi or contra hemispheres. Grey squares indicate animal medians. KW, χ2(3) = 110.685, p < 0.0001, pairwise, **p < 0.005 from Ctl. (E) Proportion of synchronous pairs in ipsi or contra hemispheres (calculated as a percentage of total pairs sampled in each hemisphere); grey squares indicate proportion of synchronous pairs for each mouse. ANOVA, F(3) = 1.297, p = 0.300. (F) Mean fractional β power (power in β relative to power 1–100 Hz) at each stage of depletion. Grey squares indicate animal means. ANOVA, p = 0.4997. (G-L). Same as A-F but for asymmetrically depleted mice. I: KW, χ2(3) = 55.317, p < 0.0001, pairwise, **p < 0.005 from Ctl. J: KW, χ2(3) = 119.892, p < 0.0001, pairwise, **p < 0.005 from Ctl. K. KW, χ2(3) = 8.782, p = 0.032, *p < 0.05 from Ctl. L. ANOVA, p = 0.6660. (M) Top: Classifier trained to predict the probability that a neuron belongs to G5%, A5%, Ipsiuni, Ipsiasym groups using: FR, firing rate; IR, irregularity; bursting, synchrony, and all pair-wise multiplicative interactions (‘ETC.’). Bottom: predicted probabilities of group membership were summed across neurons from each held-out mouse (jack-knife), and thresholded using Winner-Take-All criterion (WTA). See: Materials and methods. (N) Average held-out mouse classification performance for G5%, A5%, Ipsiuni and Ipsiasym mice (‘4-way class’) using single unit physiology. Right-tail t-test vs. chance (25%), p = 0.1986, n = 33 mice. Grey squares indicate classification accuracy of individual held out mice. Inset text indicates mice predicted >chance/ total mice. (O) Same as M but for classifying Ctl from combined Ipsiuni and Ipsiasym mice (‘Ipsi’). Right-tail t-test vs. chance (50%), p = 3.59 × 10−6, n = 23 mice, ***p < 0.001. (P) Same as M but for classifying Ctl from all end-stage mice combined (‘ES’ = Ipsiuni, Ipsiasym, G5%, A5%). Right-tail t-test vs. chance (50%), p = 2.44 × 10−5, n = 40 mice, ***p < 0.001.