Fig 1.
Schematic of experimental design.
A. Schematic representation of injection sites in a coronal mouse brain section. Sepw1-Cre mice preferentially express Cre recombinase in striatal patches and ‘exo-patches’ (see text; green). AAV5-AAV-flex-taCasp3-TEVp (0.5 μl) or sterile PBS (control) was injected bilaterally into the dorsal striatum of Sepw1-Cre mice to preferentially lesion patches. B. Representative image of intact striatum of Sepw1-Cre X Rosa26-EGFP mice displaying dense GFP expression in striatal patches. Dotted line denotes border of the striatum and solid white line denotes striatal patches. C. In lesioned mice, GFP + cells are greatly reduced and striatal patches are reduced in number. D-F. Representative μ-opioid receptor (D) and GFP expression (E) and overlay (F) in intact striatum. G-I. Representative μ-opioid receptor (G) and GFP expression (H) and overlay (I) in lesioned striatum. J. Experimental design. Mice were trained to respond on a continuous reinforcement training (CRF) before beginning variable interval 30 training (VI30). This was followed by variable interval 60 (VI60) training to establish habitual responding. After training, mice experienced counterbalanced valuation/devaluation probes (Val, Deval, respectively), followed by a day of reinstatement (VI60), and two days of omission (Omis). See Methods for details of each behavioral schedule.
Fig 2.
Lesioning striatal patches increases response variability.
A. Across CRF (FR1), variable interval 30 (VI30), and variable interval 60 (VI60) training, lesion (red) and non-lesion control (blue) mice have similar increases in press rates. B-C. Representative day-to-day variation of press rates for a control (B) and lesioned (C) mouse. The line at 0 represents the mean press rate across all VI60 days for each respective mouse and bars represent the difference from the mean on each day. D. Coefficient of variation in press rates across VI60 training days is significantly increased in lesioned mice relative to controls. E. Autocorrelation coefficient at lag 1 is reduced in patch lesioned mice relative to controls. F. Press rates across all VI60 days are not different between lesioned and control mice. * indicates p < 0.05.
Fig 3.
Lesioned mice develop a more efficient behavioral strategy.
A-B. Distribution of inter-press interval for lesioned (A) and control mice (B) on VI60 day 1 and day 5. Solid lines represent mean and dotted lines of the same color are SEM. *indicates significantly different bins based on Kolmogorov-Smirnov test (Dvalue>Dcritical) C. Lesioned mice become more efficient (change in # presses / # rewards) across training, while controls do not. D-E. Distribution of inter-entry-interval for lesioned (D) and control mice (E) on VI60 day 1 and day 5. Solid lines represent mean and dotted lines of the same color are SEM. *indicates significantly different bins based on Kolmogorov-Smirnov test (Dvalue>Dcritical) F. Lesioned mice become slightly more efficient (# head-entries / # rewards) across training, while controls do not. G-H. Cross-correlation of press rate and head entry rate in 100 ms bins for control (G) and lesioned (H) mice (lags -50 to 50; see text for details). # indicates p < 0.1; * indicates p < 0.05.
Fig 4.
Lesions of striatal patches do not change response rate during devaluation.
A-B. Press rates do not differ between devaluation and valuation probes for control (A) or lesioned mice (B). C. Press rates do not differ between lesioned (red) and control (blue) mice in devaluation D. Devaluation press rates normalized baseline press rate (average VI60 rate) is not different between lesion and control mice. E. Devaluation press rates normalized to valuation press rates (LPr, see text) did not differ between lesioned and control mice. F. Lesioned and control mice both decrease response rates across subsequent probe days.
Fig 5.
Lesions of striatal patches enhance variance across probes and alter performance in omission.
A. Variance in press rate across devaluation/valuation probe, reinstatement, and omission day 1 suggest caspase lesions increase variability in press rates across days. B. During reinstatement to a VI60 schedule following devaluation/valuation probes, mean press rates do not differ between groups. C. When controlling for differences in baseline press rate, lesioned mice increase responding to a greater extent than controls during reinstatement to the VI60 schedule (data normalized to final day of VI60). D. Mice then underwent omission across two days. Mean press rates did not differ between control and lesioned mice in either day of omission. E. When controlling for differences in baseline press rate, lesioned mice rapidly reduce press rates relative to controls in day 1 of omission (data normalized to VI60 reinstatement rates). F. Press rates within the first and second half of omission day 1 suggest reduced responding in lesion mice relative to control mice. G-H. There is a significant correlation between average VI60 press rates and press rates in the first day of omission for control (G) but not lesioned mice (H). I. One sample t-test comparing change in response from reinstatement to omission suggests that control animals display habitual responding across days (left), while lesioned animals do not (right). * indicates p < 0.05.
Fig 6.
Patch lesions slightly impair motor learning but not overall performance or locomotion.
A. Performance on the rotarod across days does not differ between lesioned and control mice. B. Performance during day 1 of rotarod training suggests that lesioned mice are slightly impaired relative to control mice. C. Performance on each trial within day 4 of rotarod training suggests that lesioned mice perform similarly late in training. D-F. Performance in open field suggests that lesioned mice are not different than controls in the overall distance moved (D), overall velocity (E), or in number of rotations (F). # indicates p < 0.1.