I'm a PhD student studying the development of cellular-scale brain networks in vitro. I grow primary murine neuronal cultures and record spontaneous activity on microelectrode arrays (MEAs). I also analyse MEA recordings of stem-cell-derived human tissue including cerebral organoid slices. I'm interested in network formation, topology and dynamics and currently focus mainly on graph theoretical metrics. As part of my PhD, I am investigating how networks containing Mecp2-deficient neurons are disrupted. This may have translational implications as Mecp2 mutations cause most cases of Rett Syndtome. Using optognetic manipulation, I am also looking at the role of inhibitory neocortical interneurons in these networks as pathology in these cells is heavily linked to Rett Syndrome.
Simulation of cortical network activity using Izhikevich (2003) model.
(a) Izhikevich (2003) model equations—v corresponds to membrane potential whereas u is a recovery variable which serves to elicit a refractory period corresponding to when K+ channels activate and Na+ channels inactive after an action potential. Four parameters impact neuronal membrane potential directly (parameter c) or indirectly via u (parameters a, b and d). Larger values of parameter a increase the rate at which the neuron returns to baseline. Larger values of b increase the sensitivity of u to a change in v meaning the neuron resets more quickly increasing likelihood of oscillations. More negative values of c increase the magnitude of post-spike hyperpolarisation thus decreasing subsequent excitability. Greater values of d increase the extent to which the recovery variable, u drops membrane potential (i.e., indirectly lowers post-spike membrane potential and slows recovery time). (b) Following a spike, defined as membrane potential, v (t) (v corresponds to millivolt scale, t corresponds to millisecond scale, Izhikevich, 2003), membrane potential is reset according to the value of c and the recovery variable, u, is also reset to the sum of its previous value and parameter d. (c) Panel from Fig. 2 of Izhikevich (2003) illustrating the two variables modelled over time, v and u and how they are impacted by the four parameters, a, b, c, and d. (d) Two types of neurons were simulated using the Izhikevich (2003) spiking model—regular spiking (RS, excitatory) and fast spiking (FS, inhibitory). RS neurons parameter settings are shown by the red dot and FS neurons parameter settings by the blue dot. These settings are based on Izhikevich (2003). (e) The membrane potential over time, v (t), equivalent to mV, for an example RS and FS neuron. (f) In order to create a network of Izhikevich neurons, they are connected by a random synaptic weight matrix with positive and negative values bounded at a magnitude of 1 reflecting excitatory and inhibitory connections, respectively. The input variable, I, of a given neuron at time, t, is the sum of synaptic weights of its inputs with v (t)>30, i.e., firing. 100 s of simulation is depicted as a raster plot with inhibitory neurons in blue and excitatory neurons in red. The corresponding distribution of firing rates of each cell type is shown to the right.
