Educational software (apps) can improve science education by providing an interactive way of learning about complicated topics that are hard to explain with text and static illustrations. However, few educational apps are available for simulation of neural networks. Here, we describe an educational app, Neuronify, allowing the user to easily create and explore neural networks in a plug-and-play simulation environment. The user can pick network elements with adjustable parameters from a menu, i.e., synaptically connected neurons modelled as integrate-and-fire neurons and various stimulators (current sources, spike generators, visual, and touch) and recording devices (voltmeter, spike detector, and loudspeaker). We aim to provide a low entry point to simulation-based neuroscience by allowing students with no programming experience to create and simulate neural networks. To facilitate the use of Neuronify in teaching, a set of premade common network motifs is provided, performing functions such as input summation, gain control by inhibition, and detection of direction of stimulus movement. Neuronify is developed in C++ and QML using the cross-platform application framework Qt and runs on smart phones (Android, iOS) and tablet computers as well personal computers (Windows, Mac, Linux).

Perineuronal nets (PNNs) are specialized extracellular matrix (ECM) structures that condense around the soma and proximal dendrites of subpopulations of neurons. Emerging evidence suggests that they are involved in regulating brain plasticity. However, the expression of PNNs varies between and within brain areas. A lack of quantitative studies describing the distribution and cell-specificity of PNNs makes it difficult to reveal the functional roles of PNNs. In the current study, we examine the distribution of PNNs and the identity of PNN-enwrapped neurons in three brain areas with different cognitive functions: the dorsal hippocampus, medial entorhinal cortex (mEC) and primary visual cortex (V1). We compared rats and mice as knowledge from these species are often intermingled. The most abundant expression of PNNs was found in the mEC and V1, while dorsal hippocampus showed strikingly low levels of PNNs, apart from dense expression in the CA2 region. In hippocampus we also found apparent species differences in expression of PNNs. While we confirm that the PNNs enwrap parvalbumin-expressing (PV+) neurons in V1, we found that they mainly colocalize with excitatory CamKII-expressing neurons in CA2. In mEC, we demonstrate that in addition to PV+ cells, the PNNs colocalize with reelin-expressing stellate cells. We also show that the maturation of PNNs in mEC coincides with the formation of grid cell pattern, while PV+ cells, unlike in other cortical areas, are present from early postnatal development. Finally, we demonstrate considerable effects on the number of PSD-95-gephyrin puncta after enzymatic removal of PNNs.

Brain research investigating electrical activity within neural tissue is producing an increasing amount of physiological data including local field potentials (LFPs) obtained via extracellular in vivo and in vitro recordings. In order to correctly interpret such electrophysiological data, it is vital to adequately understand the electrical properties of neural tissue itself. An ongoing controversy in the field of neuroscience is whether such frequency-dependent effects bias LFP recordings and affect the proper interpretation of the signal. On macroscopic scales and with large injected currents, previous studies have found various grades of frequency dependence of cortical tissue, ranging from negligible to strong, within the frequency band typically considered relevant for neuroscience (less than a few thousand hertz). Here, we performed a detailed investigation of the frequency dependence of the conductivity within cortical tissue at microscopic distances using small current amplitudes within the typical (neuro)physiological micrometer and sub-nanoampere range. We investigated the propagation of LFPs, induced by extracellular electrical current injections via patch-pipettes, in acute rat brain slice preparations containing the somatosensory cortex in vitro using multielectrode arrays. Based on our data, we determined the cortical tissue conductivity over a 100-fold increase in signal frequency (5-500 Hz). Our results imply at most very weak frequency-dependent effects within the frequency range of physiological LFPs. Using biophysical modeling, we estimated the impact of different putative impedance spectra. Our results indicate that frequency dependencies of the order measured here and in most other studies have negligible impact on the typical analysis and modeling of LFP signals from extracellular brain recordings.

The activity pattern and temporal dynamics within and between neuron ensembles are essential features of information processing and believed to be profoundly affected by anesthesia. Much of our general understanding of sensory information processing, including computational models aimed at mathematically simulating sensory information processing, rely on parameters derived from recordings conducted on animals under anesthesia. Due to the high variety of neuronal subtypes in the brain, population-based estimates of the impact of anesthesia may conceal unit- or ensemble-specific effects of the transition between states. Using chronically implanted tetrodes into primary visual cortex (V1) of rats, we conducted extracellular recordings of single units and followed the same cell ensembles in the awake and anesthetized states. We found that the transition from wakefulness to anesthesia involves unpredictable changes in temporal response characteristics. The latency of single-unit responses to visual stimulation was delayed in anesthesia, with large individual variations between units. Pair-wise correlations between units increased under anesthesia, indicating more synchronized activity. Further, the units within an ensemble show reproducible temporal activity patterns in response to visual stimuli that is changed between states, suggesting state-dependent sequences of activity. The current dataset, with recordings from the same neural ensembles across states, is well suited for validating and testing computational network models. This can lead to testable predictions, bring a deeper understanding of the experimental findings and improve models of neural information processing. Here, we exemplify such a workflow using a Brunel network model.