Drift and stabilization of cortical response selectivity

by Alexander Schmidt
Doctoral thesis
Date of Examination:2020-11-23
Date of issue:2021-08-02
Advisor:Prof. Dr. Fred Wolf
Referee:Prof. Dr. Fred Wolf
Referee:Prof. Dr. Siegrid Loewel
Referee:Prof. Dr. Jörg Enderlein
Referee:Prof. Dr. Florentin Wörgötter
Referee:Prof. Dr. Nils Brose
Referee:Dr. Alexander Becker
Persistent Address: http://dx.doi.org/10.53846/goediss-8733

 

 

Files in this item

Name:AlexanderSchmidt_PhDThesis.pdf
Size:22.1Mb
Format:PDF

The following license files are associated with this item:

Abstract

English

Synaptic turnover and long term functional stability are two seemingly contradicting features of neuronal networks, which show varying expressions across different brain regions. Recent studies have shown, how both of these are strongly expressed in the hippocampus, raising the question how this can be reconciled within a biological network. In this work, I use a data set of neuron activity from mice behaving within a virtual environment recorded over up to several months to extend and develop methods, showing how the activity of hundreds of neurons per individual animal can be reliably tracked and characterized. I employ these methods to analyze network- and individual neuron behavior during the initial formation of a place map from the activity of individual place cells while the animal learns to navigate in a new environment, as well as during the condition of a constant environment over several weeks. In a published study included in this work, we find that map formation is driven by selective stabilization of place cells coding for salient regions, with distinct characteristics for neurons coding for landmark, reward, or other locations. Strikingly, we find that in mice lacking Shank2, an autism spectrum disorder (ASD)-linked gene encoding an excitatory postsynaptic scaffold protein, a characteristic overrepresentation of visual landmarks is missing while the overrepresentation of reward location remains intact, suggesting different underlying mechanisms in the stabilization. In the condition of a constant environment, I find how turnover dynamics largely decouple from the location of a place field and are governed by a strong decorrelation of population activity on short time scales (hours to days), followed by long-lasting correlations (days to months) above chance level. In agreement with earlier studies, I find a slow, constant drift in the population of active neurons, while – contrary to earlier results – place fields within the active population are assumed approximately randomly. Place field movement across days is governed by periods of stability around an anchor position, interrupted by random, long-range relocation. The data does not suggest the existence of populations of neurons showing distinct properties of stability, but rather shows a continuous range from highly unstable to very stable functional- and non-functional activity. Average timescales of reliable contributions to the neural code are on the order of few days, in agreement with earlier reported timescales of synaptic turnover in the hippocampus.
Keywords: Neurosciences; Hippocampus; Data analysis; Place cells; Stability; Plasticity; Imaging
 
Statistik