Final Report Summary - BASALGANGLIANETWORKS (Basal ganglia and the control of locomotion)
The main objective of this project was to understand how neuronal populations in the basal ganglia contribute to the control of locomotion and the execution of habits. Reaching this objective implied to optimise two challenging technical aspects. One the one hand, we needed to establish a behavioural paradigm in which rats would learn to repetitively perform a motor skill. Importantly for our question, this paradigm had to allow an exhaustive and precise quantification of the kinematics of the movements performed during skill execution. We therefore designed a running task in a treadmill. In short, to obtain reward and optimise their efforts, rats were trained to control their running speed in a timely, precise, and stereotyped manner. This original behaviour paradigm combines two main advantages. Firstly, it focuses on a simple and natural motor command which is well known to be under the control of the basal ganglia: the locomotion. Secondly, a treadmill allows the use of lateral cameras and sensors (like accelerometers) that can fully capture the dynamics of locomotion over several seconds.
On the other hand, we needed to use electrophysiological tools capable of recording simultaneously the spiking activity of ensemble of striatal neurons. Using low impedance custom-made tetrodes array and Neuralynx NLX9 microdrive we succeeded in simultaneously recording up to 80 striatal neurons in a freely moving animal.
The main results derived from those major technical efforts are as follows: Rats were trained to run on a treadmill that can be turned off by the animals entrance in a stop area. Importantly, to turn the treadmill off, animals had to run outside the stop area for a duration longer than a fixed time interval (7 s for rats 1-2 and 5 s for rat 3) since the starting of the treadmill (correct trials, rewarded by sucrose solution). Entering the stop area before the fixed time interval triggered a warning sound and forced the rats to run for 20 s (incorrect trials, not rewarded). Naive animals instinctively rushed forward when the treadmill was turned on, which resulted in early stop area entrances and a high proportion of non-rewarded 20 s long trials (upper panels). Consequently, to perform correct trials (i.e. to stop the treadmill, run less and receive rewards), animals had to control their running speed in a timely manner. In a long trial-and-error process, animals achieved such control by developing a running routine that could be broadly divided in three phases:
1. passive displacement toward the treadmill end;
2. steady run at the end portion of the treadmill; and
3. progressive acceleration toward stop area. Consistent with a habit-based behaviour, this routine was highly stereotyped once learning was achieved and persisted for several days when the task structure was modified by shortening the treadmill length.
After this behaviour was over-trained, we investigate how ensemble of neurons in the DLS contribute to the execution of habits. Each recording session consisted in 50 to 100 trials. Tetrode recordings were obtained when animals displayed highly habitual running pattern (at least 8 weeks of training and running routine persisted when the treadmill length was shortened). Two-thirds of the recorded units had their spiking activity task-modulated from which approximately 10 % had a firing pattern compatible with gating function and fired before and after the runs. The other 90 % of the task-modulated neurons fired at different timed during the runs. Strikingly, sorting in time the normalised perivent histograms of all modulated units revealed that the activity of DLS neurons is sequentially modulated during the entire task. This suggests that DLS ensembles are involved in the ongoing execution of this habit, rather than its gating. To confirm this possibility, we examined the relationship between, spiking activity in the DLS and the locomotor dynamics and kinematics of the habit. Firstly, we found that about 10 % of the task-modulated neurons fired in a phase-locked manner with the locomotor limb movements. Secondly, using partial correlation and multiple regression analysis, we found that the firing rate of a large fraction of the task-modulated units (> 80 %) significantly correlated with running acceleration, speed, position of the animal, time or a combination of these kinematic variables.
On the other hand, we needed to use electrophysiological tools capable of recording simultaneously the spiking activity of ensemble of striatal neurons. Using low impedance custom-made tetrodes array and Neuralynx NLX9 microdrive we succeeded in simultaneously recording up to 80 striatal neurons in a freely moving animal.
The main results derived from those major technical efforts are as follows: Rats were trained to run on a treadmill that can be turned off by the animals entrance in a stop area. Importantly, to turn the treadmill off, animals had to run outside the stop area for a duration longer than a fixed time interval (7 s for rats 1-2 and 5 s for rat 3) since the starting of the treadmill (correct trials, rewarded by sucrose solution). Entering the stop area before the fixed time interval triggered a warning sound and forced the rats to run for 20 s (incorrect trials, not rewarded). Naive animals instinctively rushed forward when the treadmill was turned on, which resulted in early stop area entrances and a high proportion of non-rewarded 20 s long trials (upper panels). Consequently, to perform correct trials (i.e. to stop the treadmill, run less and receive rewards), animals had to control their running speed in a timely manner. In a long trial-and-error process, animals achieved such control by developing a running routine that could be broadly divided in three phases:
1. passive displacement toward the treadmill end;
2. steady run at the end portion of the treadmill; and
3. progressive acceleration toward stop area. Consistent with a habit-based behaviour, this routine was highly stereotyped once learning was achieved and persisted for several days when the task structure was modified by shortening the treadmill length.
After this behaviour was over-trained, we investigate how ensemble of neurons in the DLS contribute to the execution of habits. Each recording session consisted in 50 to 100 trials. Tetrode recordings were obtained when animals displayed highly habitual running pattern (at least 8 weeks of training and running routine persisted when the treadmill length was shortened). Two-thirds of the recorded units had their spiking activity task-modulated from which approximately 10 % had a firing pattern compatible with gating function and fired before and after the runs. The other 90 % of the task-modulated neurons fired at different timed during the runs. Strikingly, sorting in time the normalised perivent histograms of all modulated units revealed that the activity of DLS neurons is sequentially modulated during the entire task. This suggests that DLS ensembles are involved in the ongoing execution of this habit, rather than its gating. To confirm this possibility, we examined the relationship between, spiking activity in the DLS and the locomotor dynamics and kinematics of the habit. Firstly, we found that about 10 % of the task-modulated neurons fired in a phase-locked manner with the locomotor limb movements. Secondly, using partial correlation and multiple regression analysis, we found that the firing rate of a large fraction of the task-modulated units (> 80 %) significantly correlated with running acceleration, speed, position of the animal, time or a combination of these kinematic variables.
