ABSTRACT In Parkinson's disease (PD), degeneration of dopaminergic (DA) neurons leads to profound motor impairment. Although motor symptom is initially treatable by the DA precursor levodopa (L-DOPA), patients experience dis- abling motor fluctuations, including a shortened duration of action for L-DOPA, only partially treated with pharmacological means and deep brain stimulation. Preventing L-DOPA's declining effectiveness will greatly improve patients' quality of life and reduce social cost. Emergence of disabling motor fluctuation is associated with the decline of a component of L-DOPA's antiparkinsonian response, known as the long duration response (LDR). The LDR is a long-lasting motor improvement that persists long after L-DOPA plasma level has re- turned to baseline, gradually decaying over many hours to days after discontinuation of L-DOPA. Motor fluctuation may be caused by LDR declining too rapidly, and treatments that halt LDR decay may prevent mo- tor fluctuations. However, the mechanism underlying LDR is currently unknown. Using two distinct motor tasks, we recently found that both induction and decay of LDR is task-specific, requiring the pairing of task exposure with L-DOPA (for LDR induction) or with L-DOPA withdrawal (for LDR decay). These results point to associa- tive learning and neuroplasticity as the underlying mechanism. Furthermore, indirect pathway medium spiny neurons (iMSNs) are activated by LDR decay, and D2 receptor (D2R) knockout greatly slowed LDR decay. Based on the above results and previous findings that i) iMSN activation suppresses movement and may nor- mally function to inhibit competing responses; ii) iMSNs undergo aberrant long-term potentiation (LTP) when DA depleted, we will test the hypothesis that gradual motor impairment during LDR decay results from aberrant LTP in specific ensembles of iMSNs that are normally suppressed during normal movement by D2R stimula- tion, but become pathologically active during task exposure if DA is depleted. Using Drd2-EGFP mice to label iMSNs, we will first examine whether L-DOPA-rescued motor performance vs. LDR decay activate different iMSN ensembles in the same mouse: we will tag task-activated iMSN ensemble at the 1st time point using a Fos-promoter driven, doxycycline-gated fluorophore, then tag task-activated iMSN ensemble at the 2nd time point using endogenous Fos labeling, and compare their co-localization. We will then examine whether this “incorrect” iMSN ensemble activated during LDR decay (visualized by a Fos-driven fluorophore) has synaptic input changes that are consistent with the occurrence of LTP. Finally, we will use Fos-driven opsins to bi- directionally modulate this “incorrect” iMSN ensemble to show its causal, pathological role: that its activation leads to task-specific motor impairment, and its inhibition recues impairment task-specifically. By demonstrat- ing the existence of, and the role of, pathological iMSN ensembles in task-specific motor impa...