To summarize the preceding sections, accurately timed activities of ensembles of corti-cal neurons are needed to excite striatal neurons above threshold whose activities have themselves to be coordinated in such a way as to exert functionally specific effects on downstream structures. Synchronous population oscillations have been proposed to be an important mechanism mediating these functions along the cortico-striatal axis. However, experimental evidence for this proposition under awake behaving conditions is sparse.
The aim of the present study was thus to investigate the patterns of oscillatory popula-tion activities present in sensorimotor cortical and striatal regions in the brains of awake organisms, their synchronization as well as their modulation by gross motor behavioral demands. In this specific constellation, it was motivated by the facts that
(i) the cortico-striatal axis as the major input stage of the cortico-basal ganglia circuitry is supposed to generally have importantintegrative functions in this brain system, a role possibly subserved by synchronous oscillations,
(ii) sensorimotor cortical areas send massive, topographically ordered projections to extended parts of functionally related striatal regions, and
(iii) classically, structures along the cortico-basal ganglia loop have been assigned a primary role in the generation and adaptation of coordinated motor behavior, and the striatum is one of the main sites of affection in Parkinson’s disease, a condition marked by severe motor symptoms.
We reasoned that if synchronous population oscillations are indeed functionally relevant for neuronal signaling along the cortico-striatal axis, then distinctly different behavioral states should be reflected in marked changes of frequency-specific synchronous oscilla-tions in a task engaging regions fundamentally important for the execution of normal movements. Furthermore, we speculated that more fine-grained changes of behavioral demand would also be accompanied by parallel modulations of synchronized oscillatory patterns of population activities within and between cortex and striatum.
To explore these issues, we established an animal model of gross motor behavior, i.e., treadmill running, and trained animals to behave reliably in the task on different speed levels. We implanted microelectrodes in sensorimotor cortical and striatal regions of well-behaving subjects for the acquisition of neuronal population activities. We analyzed the data with a focus on spectral parameters. We used power and phase-coupling measures to explore the frequency characteristics and synchronization properties of population activities along the cortico-striatal axis during resting and running states. We examined interactions between frequency-specific rhythms andinterrelations of coupling measures as well as thescaling of synchronous oscillations with changing behavioral demand.
2 Methods
2.1 Treadmill apparatus and behavioral environment
The treadmill apparatus used in this study (TSE Systems GmbH; Bad Homburg, Ger-many) was similar to many of those that have been successfully employed since decades for corresponding behavioral paradigms (e.g., Kimeldorf, 1961; Andrews, 1965; Chapin et al., 1980). We describe here its basic composition and configuration before we turn to more specific modifications made to it and the behavioral environment as prepared for recordings in a later section (2.4.2). Note that large parts of this and the following section (2.2) as well as some minor parts of Section 2.4.2 have been reused and adapted from another manuscript of the author describing in detail the establishment of a treadmill running model in Brown Norway rats (von Nicolai, 2011).
The main part of the treadmill (Figure 2.1) consisted of a rubber belt wrapped around two pulleys whose axes were fixed to an aluminum plate. The belt was stretched tightly between the two pulleys such that it served as a moving ground above the plate on which rodents could sit and walk. This main part was 50ˆ13 cm in size, surrounded by a frame of glass fibre reinforced plastic walls of about 16.5 cm height and covered with a plexiglas top that could be opened and removed for handling of the animals. Another plastic panel was mounted on top of the belt and in the middle between the two sidewalls to split the belt’s surface into two lanes for two rats to be run in parallel.
The belt setup was connected to a small, computer-driven motor operated by means of a control software installed on a PC according to user-definable settings. The inclination of the treadmill could also be changed both manually and automatically to range between 0 and 20 degrees with respect to ground level. However, this feature was not used in any of the experiments carried out during the course of this project.
At the back end of the treadmill there was a metal grid attached to the pulley for animals to sit when they had stopped running and had been carried to the back by the moving belt. Originally, this grid had been designed for the application of mild electrical shocks in order to motivate animals to keep running in the task. This shocking device was not used in the present study for several reasons. First, as the results of extensive training sessions on many subjects prior to implantations had shown, the animals used
a b
Figure 2.1: (a) Side view photograph image of the treadmill device used in the project as originally distributed by the manufacturer. Image courtesy of TSE Systems GmbH.
(b) Rear view photograph image of the treadmill running lanes as seen through the half-open plexiglass top. The picture shows the wooden walls placed right in front of the shocking grids. Also note the darker, red area at the front end of the lanes that served as an additional incentive for the animals to keep running straight.
here could not be motivated to get back to running in response to the shocks but rather gripped onto the grid and stayed there during and after the shocking intervals. Second, for electrophysiological recordings it was highly desirable to largely remove any unnatural kinds of disturbances from the recording environment that could possibly interfere with the experiments. Third, previous studies have shown that the use of a shocking device can be substantially stressful for the subjects (Burgess et al., 1991). Lastly, the gap between the metal grid and the running belt was another source of potential harm for the animals since they could potentially have become entangled with their feet and tails there as also others have explicitly noted (Nakao et al., 1982).
For all these reasons, a different strategy was chosen to keep animals running straight on the track. The treadmill was modified by fixing a wooden wall at the back end of each lane right in front of the shocking grid (Figure 2.1, panel b). Foamed material was attached to their lower ends in order to protect the animals’ feet and tails from being scratched at the edges of the walls. We reasoned that animals would experience their hitting of the back wall in case of running errors as a sufficiently aversive event that would motivate them to resume running as quickly as possible. While this turned out to be the case in only a subset of animals, we feel that this approach of trying to keep animals stay on the lanes was by far much better suited to the needs and constraints of the present study and also much less risky and harmful for our subjects.
Chapter 2 Methods 2.2
Behavioral performance on the treadmill was measured in terms of breaks of infrared light beams. These were situated within the front ends of two pairs of thin metal rods attached to the inner walls closely adjacent to the belt’s surface at the back end of both lanes. Since the running belt moved from front to back, an animal that would not keep running straight towards the upper end of its lane would be carried to the back and then interrupt the beam. These running errors (beam breaks) were counted for each subject individually by the Process Control Unit that connected the treadmill to a PC.
The treadmill control software that was used to operate the machine also generated a detailed record of the number, time points, and epochs of running errors. In this way, we were able to monitor the behavior at all stages of a given running protocol. These protocols could be designed by the user according to specific needs in terms of trial and epoch lengths, speed levels, and the size of speed changes. The speed range covered by the device was limited to values between0.07and2.00m/s, and the maximum resolution of behavioral measurements was 5 seconds (i.e., a light beam break could only be elicited once every 5 seconds). This restricted both the accuracy with which the running behavior of the animals could be evaluated and the amount of behavioral control achieved with respect to changes of neuronal activities (see Chapters 3 and 4). For an illustration of the full makeup of trials used for recordings in terms of levels and their basic composition of 5-second epochs, see Figure 2.6 on page 36.