ESI Lecture by Philipp Haueis
- Sunday, July 23, 2017, 16:00-17:30
- Lecture Hall, ESI
- Philipp Haueis (Max Planck Institute for Brain and Cognitive Sciences)
Gain modulation and gain control: two functional roles of gamma oscillations
Gain modulation” refers to mechanisms by which oscillations in the gamma range (30–80Hz) can enhance neuronal firing either pre-synaptically by synchronizing inputs arriving at the dendritic tree, or post-synaptically by synchronizing input to phases of high excitability within a local microcircuit. Post-synaptic gain modulation has been recently been shown to be involved in behaviorally relevant information processing: behaving macaques showed shorter reaction times when attending to visual stimulus changes following phases of gamma oscillations that co-occurred with strong neural responses (Ni et al. 2016). These results support the “communication through coherence hypothesis” (CTC), according to which the functional role of gamma oscillations is to create temporal windows during which information flow between synaptic sites is enhanced (Fries 2015). In this talk, I propose that the information processing role of gamma oscillations in multiplicative gain modification could be supplemented by their infrastructural role in gain control (Kirschfeld 1992, Merker 2013). Because gamma oscillations result from excitatory-inhibitory interactions over time, they could provide negative feedback that signals circuit activation levels to inhibitory chandelier cells. Since chandelier cells form serial cartridge synapses with pyramidal axon hillocks, they can both facilitate or control postsynaptic gain, depending on whether the hillock membrane is below or above the EGABA reversal potential (Szabadics et al. 2006). This means that gamma oscillations could be both part of a positive feedforward mechanism that enhances behaviorally relevant information processing (cognitive function) and a negative feedback mechanism that prevents over- excitation and circuit damage through seizure activity (infrastructural function). Drawing on the analysis of mechanisms from philosophy of neuroscience (Craver 2007), I argue that these cognitive and infrastructural functions are mechanistically coupled: it is not possible to manipulate the gain modification pathway independently from the gain control pathway and vice versa. I conclude by showing that despite mechanistic coupling, the distinction between cognitive and infrastructural functions remains useful, as it sheds light on the issue of information flow within neural circuit elements.