In this thesis, we explore different topics related to information processing in neuronal circuits. Understanding information processing in biological networks requires not only understanding the information-theoretic consequences of neuronal activity but also understanding the network and single-cell dynamics and transformations underlying those responses. Therefore, we take on two different perspectives on information processing in the brain. First, we study the information-theoretic consequences of different shapes of tuning curves. Here, we depart from the traditional method of studying information through asymptotical measures such as Fisher information or mutual information and instead focus on the impact of tuning shapes and decoding times on rare but large estimation errors. We show that studying the role of decoding time reveals new interesting constraints on the "neural code." We argue that these constraints might explain the tuning organization found in early sensory systems. Second, we explore the role of single-cell and network dynamics on the input-output transfer function of spike trains. Understanding how single cells and networks of cells transform external signals is a key component in understanding the constraints and possibilities to encode information in neuronal circuits. In particular, we study the role of NMDA in expanding the post-synaptic range of sensitivity to pre-synaptic activity into states of high conductance using modeling of a morphologically reconstructed cell. We show that the NMDA-AMPA ratio can be an important mechanism to control the excitability of a cell given its natural range of inputs. Lastly, we study the joint input-output transformation of firing rate and synchrony in networks of point neurons. Synchronous inputs have been proposed to increase the pre-synaptic efficacy in electing post-synaptic responses but can also induce strong synchronization in the downstream networks. We show that feedforward inhibition, if tightly correlated with the feedforward excitation, can reduce post-synaptic synchronization at the expense of input synchrony no longer driving increased post-synaptic activity but, if anything, hinders it.

I denna avhandling utforskar vi olika ämnen kring informationsbearbetning i neurala nätverk. För att förstå informationsbearbetning i biologiska nätverk räcker det inte att endast studera de informationsvetenskapliga konsekvenserna av hjärnaktivitet, vi måste också förstå de underliggande dynamiska egenskaperna och transformationerna hos de celler och nätverk som ger upphov till aktiviteten. Därför tar vi an två perspektiv i denna avhandling. I den första delen undersöker vi, med hjälp av informationsvetenskapliga metoder, hur olika responskurvor (tuning curves) påverkar information. Vi frångår här från det vanliga användandet av asymptotiska mått, så som Fisher information eller ömsesidig information, och fokuserar istället på konsekvensen av korta avkodningstider. Vi visar att korta avkodningstider kräver andra, nya hänsynstaganden för en neural kod. Detta nya perspektiv ger en ny informationsvetenskaplig förklaring till de responskurvor som uppmäts i våra sensoriska nervsystem. I den andra delen av denna avhandling studerar vi transformationen av hjärnaktivitet i enstaka celler eller i nätverk av celler. Vi studerar en möjlig implikation av NMDA i att möjliggöra post-synaptisk sensitivitet i situationer av hög pre-synaptisk aktivitet genom simulationer av en morfologiskt rekonstruerad cell. Vi visar att den synaptiska NMDA-AMPA ration kan vara en viktig faktor med vilken celler kan anpassa sin excitabilitet efter dess naturliga pre-synaptiska aktivitetsnivå. Slutligen studerar vi även den gemensamma effekten av pre-synaptisk aktivitetsnivå och synkronicitet på den post-synaptiska aktiviteten i nätverk av punktneuroner. Synkroniserad aktivitet har föreslagits som en mekanism att öka den pre-synaptiska förmågan att driva post-synaptisk aktivitet. Dock kan synkron pre-synaptisk aktivitet även ge upphov till ytterligare synkronicitet i senare nätverk. Vi föreslår att inhibering, starkt korrelerad med den externa exciteringen, kan kraftigt reducera post-synaptisk synkronicitet. Slutligen visar vi att sådan dekorrelation även medför lägre aktivitet.

According to the efficient coding hypothesis, sensory neurons are adapted to provide maximal information about the environment, given some biophysical constraints. In early visual areas, stimulus-induced modulations of neural activity (or tunings) are predominantly single-peaked. However, periodic tuning, as exhibited by grid cells, has been linked to a significant increase in decoding performance. Does this imply that the tuning curves in early visual areas are sub-optimal? We argue that the time scale at which neurons encode information is imperative to understand the advantages of single-peaked and periodic tuning curves, respectively. Here, we show that the possibility of catastrophic (large) errors creates a trade-off between decoding time and decoding ability. We investigate how decoding time and stimulus dimensionality affect the optimal shape of tuning curves for removing catastrophic errors. In particular, we focus on the spatial periods of the tuning curves for a class of circular tuning curves. We show an overall trend for minimal decoding time to increase with increasing Fisher information, implying a trade-off between accuracy and speed. This trade-off is reinforced whenever the stimulus dimensionality is high, or there is ongoing activity. Thus, given constraints on processing speed, we present normative arguments for the existence of the single-peaked tuning organization observed in early visual areas.