diff --git a/sbBrix/common/policies.py b/sbBrix/common/policies.py
index 66de5c8..ce6a18a 100644
--- a/sbBrix/common/policies.py
+++ b/sbBrix/common/policies.py
@@ -676,6 +676,8 @@ class ActorCriticPolicy(BasePolicy):
             return self.action_dist.proba_distribution(action_logits=mean_actions)
         elif isinstance(self.action_dist, StateDependentNoiseDistribution):
             return self.action_dist.proba_distribution(mean_actions, self.log_std, latent_pi)
+        elif isinstance(self.action_dist, PCA_Distribution):
+            return self.action_dist.proba_distribution(mean_actions, self.log_std)
         else:
             raise ValueError("Invalid action distribution")