Fellow-Klasse 2010/11 Prof. Dr. Hannah Monyer

Arbeitsvorhaben am Marsilius-Kolleg

The impact of brain rhythms in the awake and sleeping rodent on spatial memory

Scientific background and own contributions

Although GABAergic neurones constitute only about a fifth of the neuronal cell population, they are endowed with functional mechanisms that ensure their control over the much larger population of principal (excitatory) neurones. GABAergic interneurones play an important
role in synchronising large neuronal ensembles giving rise to a wide range of oscillatory patterns. In the hippocampus, theta oscillations (6-10 Hz) concurrent with gamma oscillations (30-80 Hz) are predominantly present during exploratory activity and REM sleep (Vanderwolf, 1969; Buzsaki, 2002, Csicsvari et al., 2003). During period of immobility, brief episodes of fast (140-200 Hz) oscillatory patterns ("ripples") are present (O'Keefe and Nadel, 1978; Buzsaki et al., 1992).

Most importantly and relevant for this proposal are many studies from other and our own laboratory demonstrating that in-phase synchrony is brought about by GABAergic interneurones (Bartos et al., 2007). Inhibition-based oscillations have been demonstrated and studied in vitro and in vivo and computational models were instrumental in providing explanations as to how GABAergic interneurones can/might effectively synchronize principal cells (reviewed in Traub, 2004). Work from our lab demonstrated that selective genetic manipulation in GABAergic interneurones resulted in altered oscillatory activity in vitro and in vivo (Fuchs et al., 2001; 2007; Hormuzdi et al., 2001; Buhl et al., 2003; Wulff et al., 2009; Racz et al, 2009).

Theta oscillations link two ways of coding by which pyramidal cells in the hippocampus represent space, namely rate and phase coding. The seminal discovery of "place cells" in the hippocampus led to the proposal that the hippocampus functions as a cognitive map subserving navigation (O'Keefe and Dostrovski, 1971). Place cell activity is sparse, indicating that in the normal experimental environment most place cells re in a single location, i.e. the "place field" of the neuron. The location of the animal is thus represented by the collective activity of place cells. The spatially correlated activity of place cells, however, varies when referenced against the phase of the theta cycle measured in the extracellular EEG (O'Keefe and Recce, 1993). The tendency of place cells to fire at progressively earlier phases of the theta rhythm, termed "phase precession", is thought to be a critical determinant of episodic memory (Skaggs et al., 1996).

Grid cells in the medial entorhinal (mEnt) cortex are the other major class of neurones whose activity is critically involved in spatial representations (Hafting et al., 2005; Moser et al., 2008). Grid cells are activated in multiple places and the regularly spaced locations of maximal firing exhibit a grid-like pattern covering the entire environment. Questions regarding the mechanisms of theta rhythm and phase precession in the hippocampus and mEnt are not resolved yet. Also, the origin of phase precession, the relation and possibly dependence between place and grid cells are still a matter of intensive debate and research. Lesion studies provide evidence that entorhinal input to CA1 not only supports spatial but also temporal firing patterns (McNaughton et al., 1989; Kocsis et al., 1999; Brun et al., 2002), and grid cell stability is affected following hippocampal inactivation (Hafting et al., 2008).

Lesion studies first in humans and later in animal models provided evidence that the hip-pocampus and associated structures are involved in memory formation and led to the identification of multiple classes of memory (Scoville and Milner, 1957; Squire, 1992). Hippocampal and entorhinal lesions result in spatial navigation and memory impairments. Studies on genetically modifed mice provided important information on the dissociation of distinct memory systems and helped assign distinct computations to subregions of the hippocampus. Global and region-specific gene ablations have constituted an important tool that furthered our understanding regarding hippocampal functions underlying spatial memory but all modifications performed so far affected either both principal cells and interneurones or principal cells alone.

A major contribution of our lab that serves as a corner stone to this proposal is the finding that selectively modifying the activity of GABAergic interneurones has a strong impact on specific aspects of spatial memory (Fuchs et al., 2007). Using two genetic approaches we demonstrated that ecient recruitment of specific GABAergic interneurones is required for spatial memory. Moreover we could show that two hippocampus-dependent memory systems were differentially affected. Thus, in mutant mice the performance was impaired in tasks assessing a memory system that supports rapidly acquired place memory, particularly in circumstances when relevant information is changing frequently. However, the mutants were undistinguishable from control mice in spatial reference memory tasks that involve a memory system supporting incremental learning over time.

Finally, and pertinent for this proposal is the finding that memory consolidation occurs during sleep. In a series of elegant experiments, Jan Born and colleagues showed that non-REM sleep plays an important role in memory formation in humans (Marshall et al., 2006). Furthermore there is compelling evidence that hippocampal CA1 place field cells that were coactive during exploratory activity exhibited an increased tendency to fire together during subsequent sleep (Wilson et al., 1994). The activity associated with the replay at the EEG level is characterized by high frequency oscillations "ripple" activity that occurs during non-REM sleep. In a recent elegant study by Girardeau and colleagues (2009), the authors could demonstrate that interfering with ripple activity during sleep, led to impairment of spatial reference memory. The increased correlated activity of neurons during sleep is thought to underlie subsequent synaptic modifications that are believed to be the basis of memories in the brain. An interesting unresolved question is where exactly do synaptic modifications occur? Thus, although CA1 hippocampal place cells exhibit selective correlated reactivation during sleep if they have been active during prior waking period, they are most likely not the substrate of synaptic modifications since there is little direct connectivity between them. The medial entorhinal cortex, a major input to CA1 is the likely structure where experience-dependent correlated activity emerges and where long-lasting changes might occur.

Proposal

Here we propose to define the functions of GABAergic interneurones at the network and behavioural level in the hippocampus and the mEnt. This will be achieved by manipulating the activity of GABAergic interneurones in a cell-specific or region-specific manner.

GABAergic interneurone activity will be manipulated by either reducing the excitatory drive onto them or by preventing electrical coupling between them.We will use wildtype and mutant mice that will be analysed focusing on rhythmic network activity, spatial coding by principal cells and spatial memory.

Electrophysiological recording in behaving mice will be performed to address 3 questions:

• Does manipulating the activity of GABAergic interneurones in the hippocampus or mEnt cortex affect behaviour in the same manner?
• What is the relation between spatial information processing in the hippocampus/mEnt cortex and different types of spatial memory?
• Does the interference with GABAergic interneuron activity in the hippocampus/mEnt cortex affect memory acquisition in the behaving animal or memory consolidation during sleep?

Combining the results obtained in electrophysiological and behavioural experiments, we will be able to link GABAergic interneurones with critical hippocampal-mEnt system functions governing spatial navigation and memory.