Animals need to switch between motivated behaviours, like drinking, feeding or social interaction, to meet environmental availability and internal needs. However, motivational switching is rarely studied, partly due to lack measurement systems. We designed an automated extended home-cage for measuring motivational switching in mice, the Switchmaze, using open source hardware and software (Figure 1). As proof-of-concept, we show environmental manipulation and targeted brain manipulation experiments which altered motivation switching without effect on traditional parameters.
Individual animals access the Switchmaze from the home-cage and choose between entering one of two chambers containing different goal objects or returning to the home-cage. Motivational switching is measured as a ratio of switching between chambers and continuous exploitation of one chamber. This motivation switching rate (MSR) was on average 1.4 ± 0.9 (mean ± SD, n=20 animals) after continuous habituation for >7 days. To test if motivational switching was different from random, we randomly shuffled the trial sequence and re-calculated MSR 1000 times for each animal. The arising distribution encompassed the actual MSR in most cases (only 4/30 animals had MSR higher than the 99th percentile of the shuffled data). This suggests that spontaneous motivation switching is optimized to appear random, which could conceivably function to decrease the predictive information available to competitors and predators.
To test the role of key neural populations in motivational switching with a loss-of-function manipulation, we expressed the inhibitory chemogenetic construct, hM4Di in either prefrontal cortex (PFC) output neurons to the hypothalamus (PFC-hM4Di) or in the perifornical hypothalamus (H-hM4Di). As a control cohort, we used wild-type mice that did not express a transgene, and which were interleaved in groups of hM4Di expressing mice that lived in the Switchmaze. We measured behaviour for 6 hours following a vehicle injection (i.p. saline) or C21, which activates hM4Di. Basic behavioural metrics were not changed in PFC-hM4Di or H-hM4Di mice upon activation of the inhibitory DREADDs with C21, as two-way repeated measures ANOVA tests showed no significant cohort-time interaction for food consumed (F(2,27)=0.61, p=0.55), block count (F(2,27)=0.66, p=0.52), trial count (F(2,27)=0.50, p=0.61), block duration (F(2,27)=0.75, p=0.48) or trial duration (F(2,27)=0.69, p=0.51). However, MSR was altered significantly (cohort-time interaction F(2,27)=3.99, p=0.03) and paired t-tests revealed the effect was a 46.5 ± 45.5% increase in the PFC-hM4Di cohort (p=0.007). The average switch rate exceeded the 99th percentile of switch rates expected from randomly permuted trial sequences.
Our results suggests that a behavioural role of the PFC is to regulate motivational switching in order to appear less predictable. The Switchmaze may be useful in scoring behavioural rigidity, which is a hallmark of many neuropsychiatric disorders. This work demonstrates the utility of open-design in understanding animal behaviour and its neural correlates.
