To study actions is to study the way we do things, which is different than studying how we remember stimuli, or facts and events. Some actions are innate or prewired (such as swallowing or breathing). Others are learned anew throughout life, likely through a process of trial and feedback. We currently focus on understanding the processes mediating the latter.
Our overall goal is to understand how changes in molecular networks in the brain modify neural circuits to allow the generation of novel actions and their shaping by experience. To achieve this, we subdivided our experiments into different subgoals:
- Action generation: focusing on the mechanisms underlying the generation of novel/diverse actions (trial),
- Action shaping and automatization: focusing on understanding the mechanisms underlying changes in how actions are performed, for example, how we improve the accuracy and speed of actions and how we organize actions into precise sequences (through trial and feedback), and
- Action goals: focusing on understanding the mechanisms underlying changes in why actions are performed, for example, how we learn that particular actions lead to particular outcomes (goal of the action) and perform them in an intentional manner, or how we form habits or routines.
A growing body of evidence suggests that cortico-basal ganglia circuits are involved in action generation and selection, in skill learning, and in learning goal-directed actions and habits. Therefore, we investigate the cortico-basal ganglia mechanisms underlying the processes described in the different subgoals by using an across-level approach, from molecules to circuits.
We use mice in this integrative approach because they combine the power of genetics, a mammalian brain with canonical cortico-basal ganglia loops that can generate and propagate oscillatory activity, and the possibility of accurately quantifying simple behaviors such as action initiation (using EMG recordings or inertial sensors) and stereotypic skill learning as well as more elaborate behaviors such as goal-directed actions.
We and other animals have the unique capacity to move and behave, change the environment around us, and develop sophisticated action skills to communicate and seek resources. Without this capacity, we cannot nourish ourselves or interact with the environment or others. We cannot play music, create, or build. In summary, we cannot be actors in the world that surrounds us.
Many of the actions we perform are innate, that is, they are genetically determined and prewired. Although innate actions may be modulated by experience, they are essentially fixed action patterns controlled by either central pattern generators or stimulus-response circuits such as reflexes (input-output). However, how do we generate novel actions—those that were never generated before and may never be generated again? The first time we produce a particular complex movement, especially during early development (motor babbling), it is unlikely that it is generated by reflexive circuits or triggered by particular stimulus-response associations.
Our lab has been investigating the hypothesis that neural circuits, and in particular cortical-basal ganglia circuits, can be active generators of variable neural activity patterns that produce novel movements. These novel actions could then be selected by interactions with the environment, to a point where they could be executed very rapidly and precisely, would be less variable, and could be elicited by particular stimuli. So, how does action generation occur? Our lab has been exploring the possibility that dopamine, which is essential for the generation of novel actions and is depleted in patients with Parkinsons disease, is a critical modulator of variability in cortico-basal ganglia circuits. According to this view, the locking of cortico-basal ganglia circuits into certain patterns of activity in the absence of dopamine would be mechanistically related to the inability to perform new voluntary movements.
We have shown that increased dopamine levels lead to more variability in corticostriatal circuit activity, whereas dopamine depletion, which causes lack of novel voluntary movement, leads to dramatic synchrony and low variability. Currently, we are investigating how different dopamine circuits/populations contribute to movement generation and variability in cortico-basal ganglia networks.
Action Shaping and Automatization
A corollary of the hypothesis presented above is that as we shape how we do an action, and automatize its execution, crystallization or consolidation of particular activity patterns should occur, and so variability in neural activity between different trials/executions should decrease. Our lab has found that when we learn a novel skill, variability in cortico-basal ganglia activity from trial to trial is high, but as we consolidate the skill, variability diminishes dramatically and activity patterns remain stable across trials. Furthermore, we have found that the decrease in variability and the stabilization of activity patterns during skill consolidation are accompanied by synaptic plasticity in basal ganglia, which renders the movement less dopamine dependent, and may explain why patients with Parkinsons can still perform well-learned skills.
However, the initiation and termination of actions that have been crystallized into precise movement sequences and can be automatically executed are still in many cases self-paced by the individual. Therefore, although those with Parkinsons (and Huntingtons) can execute well-learned movements, they still have difficulties in initiating and terminating them. Our lab has discovered that as we learn and automatize novel action sequences, activity specifically related to either the initiation or the termination of action sequences develops in nigrostriatal circuits (dopaminergic, striatal, and reticulata neurons). Currently, using optogenetics and increasingly complex sequence tasks, our lab is dissecting the involvement of different cell types in cortico-basal ganglia in the learning and execution of action sequences.
Even after automatization of a particular action or movement sequence, we have to decide to execute it or not and when to execute it. Hence, our lab is also investigating why we perform a particular action. What drives us? Is it the outcome, that is, the goal? Or is it a particular situation that triggers the action automatically?
There is increasing evidence that initially, as we learn that performing a novel action leads to a specific outcome, the action is mostly goal directed, that is, its performance is sensitive to changes in the value of the outcome and to changes in the contingency between obtaining the outcome and performing the action. With repetition, however, actions can become not only more efficient but also more habitual, that is, less sensitive to changes in contingency or outcome value and more controlled by antecedent stimuli.
Our lab has confirmed that in mice, as in rats and primates, the neuroanatomical circuits supporting the learning and performance of goal-directed actions are different from those supporting the formation of habits. Associative cortico-basal ganglia circuits involving the dorsomedial striatum are important for goal-directed behavior, whereas the dorsolateral or sensorimotor striatum controls habit formation. Additionally, using both genetically targeted mice and pharmacology, our lab has found that endocannabinoid signaling through CB1 receptors, which are highly expressed in the sensorimotor striatum, is critical for habit formation. Furthermore, using cre-mediated deletion of NMDA receptors in different midbrain dopamine neurons, we discovered that nigral but not mesolimbic dopamine seems to be important for habit formation.
Forming habits is important, but in our everyday life, we constantly encounter circumstances that cause us to reevaluate the consequences of our actions. Therefore, shifting between habits and goal-directed actions allows us to act flexibly and efficiently in our environment. Currently, our lab is investigating the mechanisms that allow us to shift between goal-directed actions and habits.
Grants from the European Research Council and the Portuguese Foundation for Science and Technology provided partial support for these projects.
As of January 17, 2012