Dysfunction in primate dorsolateral prefrontal area 46 affects motivation and anxiety

Abstract

The dorsolateral prefrontal cortex (dlPFC) is a higher-order brain structure targeted for noninvasive stimulation for treatment-resistant depression. Nonetheless, its causal role in emotion regulation is unknown. We discovered that inactivating dlPFC area 46 in marmosets blunts appetitive motivation and heightens threat reactivity. The effects were asymmetric-dependent on the left hemisphere only-and were mediated through projections to pregenual cingulate area 32. The antidepressant ketamine blocked the appetitive motivational deficits through mechanisms within subcallosal cingulate area 25, an area linked with treatment success in dlPFC noninvasive stimulation. Our data uncover an integrated prefrontal network for area 46 that contributes to positive and negative emotion regulation that may be core to our understanding of symptoms and therapeutic strategies for treatment-resistant depression and anxiety.

Fig. 1. Chemogenetic inactivation of A46 blunts appetitive motivation that can be ameliorated by ketamine treatment systemically and by direct A25 infusion.

(A) Schematic of hM4Di viral construct with photomicrograph of fused HA-tag expression in A46 with inset coronal section showing region of interest. (B) Marmosets respond to a touchscreen stimulus a progressively increasing number of times for reward in the progressive ratio (PR) task. (C) DCZ treatment decreased touchscreen responses compared to vehicle (Veh; n=6, 10μg/kg i.m. 30min before testing; paired t-test, p=0.0016, d=2.52). (D) Preference for 6% sucrose over water was unaffected by DCZ treatment (n=4, paired t-test, p=0.408). (E) Intramuscular ketamine (Ket) 24 hours prior to testing blocked the DCZ-induced reduction in responses (n=6; repeated measures ANOVA (rmANOVA) pretreat*treat interaction, F_(1,5)=15.24, p=0.011, η²=0.753; post-hoc: Sal/Veh vs Sal/DCZ, p=0.001, d=2.62; Sal/DCZ vs Ket/DCZ, p=0.002, d=2.41; Sal/Veh vs Ket/Veh, p=0.508). (F) Schematic for assessing A25’s involvement in A46-induced motivational deficits through ketamine infusion. (G) Intra-A25 ketamine infusion blocked DCZ-induced reduction in responses (n=6; rmANOVA, pretreat*treat interaction, F_(1,5)=50.34, p<0.001, η²=0.91; post-hoc: Sal/Veh vs Sal/DCZ, p<0.001, d=3.35; Sal/DCZ vs Ket/DCZ, p=0.005, d=1.99; Sal/Veh vs Ket/Veh, p=0.794). Data displayed as means ± SEM and individual data points with significant Sidak corrected post-hoc comparisons indicated by *.

Fig. 2. Inactivation of an A46 to A32 pathway reduces appetitive motivation with this effect blocked by a ketamine infusion in A25.

(A) Schematic of A46 terminal pathway manipulations through DCZ infusion (100nM) in A32, A32v or A25 of the marmoset. (B) DCZ infusion into A32 reduced the total responses marmosets made in comparison to the previous day, with DCZ into A32v and A25 showing no effect (n=6, rmANOVA, region*treatment interaction, F_{(1.73, 8.63)}=33.17, p<0.001, η²= 0.869; post-hoc: A32 Veh vs DCZ, p=0.001, d=2.77; A32 DCZ vs A32v DCZ, p=0.003, d=2.78; A32 DCZ vs A25 DCZ, p=0.004, d=2.65). (C) Schematic for assessing A25’s involvement in A46 to A32 pathway induced motivational deficits through A25 ketamine infusion. (D) Intra-A25 ketamine infusion blocked the ability of an A32 DCZ infusion from reducing responses (n=4, rmANOVA, pretreat x treat interaction, F_(1,3)=15.31, p=0.030, η²= 0.836; post-hoc: Sal/Veh vs Sal/DCZ, p=0.014, d=2.60; Sal/DCZ vs Ket/DCZ, p=0.019, d=2.32; Sal/Veh vs Ket/Veh, p=0.461). Data are displayed as mean ± SEM and individual data points with significant Sidak corrected post-hoc comparisons indicated by *.

Fig. 3. Chemogenetic inactivation of A46 heightens responsivity to ambiguous threat through projections to ventral A32.

(A) Schematic of the human intruder paradigm setup where a novel human maintains eye contact with a marmoset for 2 minutes. A difference score between treatments is used for each marmoset to account for individual variability (n=6). (B) DCZ treatment (10μg/kg, i.m.) increased the threat score compared to vehicle treatment (one sample t-test vs 0, p=0.005, d=1.98). (C) DCZ infusion into A32v alone heightened the threat score when compared to vehicle infusion, and also when compared to the effects of A32 and A25 DCZ infusion (rmANOVA, region effect, F_{(1.7, 8.5)}=12.52, p=0.004; post-hoc, A32 vs A32v, p=0.022, A32v vs A25, p=0.039; A32v vs 0, p=0.012, d=1.56). Data are displayed as means ± SEM and individual data points with significant Sidak corrected pairwise comparisons indicated by * and one sampled t-test vs hypothetical mean of 0 by #.

Fig. 4. Causal evidence for functional asymmetry across A46 hemispheres through unilateral behavioral analysis and correlatory evidence from resting state functional connectivity clustering.

(A) Bilateral and left A46 inactivation by muscimol/baclofen (MB) infusion enhanced responsivity to threat whilst right A46 inactivation had no effect (rmANOVA, region effect, F_{(1.8, 10.7)}=13.5, p=0.0014; post-hoc, Bi vs Right, p=0.013; Left vs Right, p=0.041). (B) DCZ infusion into left A32v increased threat responsivity when compared to right A32v (one sampled t-test vs 0, p=0.008). (C) Left and bilateral A32 DCZ infusions reduced responses in the PR task, with right DCZ infusion having no effect (rmANOVA, region effect, F_{(2.9, 14.3)}=42.8, p<0.001; post-hoc, Bilat Veh vs Left DCZ, p=0.003; Bi Veh vs DCZ, p=0.004; Left vs Right DCZ, p=0.003; Right vs Bilat DCZ, p=0.002). (D) K-means clustering of A46 resting state functional connectivity data in 20 subjects with k=2 indicated separate bilateral clusters that align with architectonically defined ventral (green) and dorsal A46 (blue) with high confidence, revealed by clustering probability obtained from bootstrap resampling. (E) K=4 clustering exposes A46 connectivity differences between pre-dominantly right (green) and left (blue) hemispheres with high confidence. Black outline indicates A46. Behavioral data displayed as means ± SEM and individual data points.