Fig 1.
Schematic of the computational model of hippocampus.
The model consists of 3080 pyramidal cells, with only 80 of them actively spiking, and 20 basket cells. Both 80 active pyramidal cells and 20 basket cells have excitatory synaptic input. Basket cells make gap junctions to the nearest basket cells. They send GABAergic synapses to pyramidal cells and receive feedback AMPAergic synapses from active basket cells. The total network spans an area of 1 × 1 mm2. Activated pyramidal cells and basket cells are distributed across a 2D plane with an area of 400 × 400 μm2 while the inactive pyramidal cells cells are distributed across an area of 1 × 1 mm2. The schematic shows the side view of the 2D array to clarify the synaptic connectivity. Recording electrodes of various sizes are placed at varying distances from the network.
Fig 2.
Synaptic potentials and neuronal spikes contributions to SPW-Rs.
(a) Spike raster for 80 active pyramidal cells (0–80) and basket cells (80–100) showing synchronous firing (for a noise intensity of 0.77 nA2). Spike raster does not show 3000 non-firing pyramidal cells. (b) LFP waveforms for total SPW-R (black), contributions from IPSPs (red), and contributions from spikes (blue) are shown. (c) Spectrograms are shown for total SPW-Rs (left), spikes (center) and IPSPs (right). IPSPs are responsible for high amplitude sharp waves (c, right). Spikes from individual neurons only contribute to ripples (c, center) at 200 Hz and weaker contributions at 300 Hz and 400 Hz. (d) Spike raster showing only basket cell firing (for a noise intensity of 2.5 × 10−4 nA2). (e) LFP waveforms for total SPW-R (black), contributions from IPSPs (red), and contributions from spikes (blue) are shown. IPSPs are the main contributor to SPW-Rs. (f) Spectrograms are shown for total SPW-Rs (left), spikes (center) and IPSPs (right).
Fig 3.
Effect of number of inactive pyramidal cells on SPW-Rs.
(a) Spectrograms of total SPW-Rs and contribution from IPSPs and LFP waveforms for total SPW-Rs (black), contributions of spikes (blue), and contributions of IPSPs (red) for 10, 100, and 1000 inactive pyramidal cells. (b-c) Graphs of the amplitudes of sharp waves (b) and ripples (c) vs. inactive pyramidal cell number.
Fig 4.
Theoretical analysis of electrode distance.
(a) Figure shows SPW-R waveforms and spectrograms calculated at 0 μm, 50 μm, and 300 μm distances from the network for point electrodes. (b) Amplitude of sharp waves is plotted as a function of distance for three different sizes of electrodes: point, 300 × 300 μm2, and 1 × 1 mm2. Sharp wave amplitude decreases significantly with increasing distance. (c) Amplitude of ripples is plotted as a function of distance. Ripple amplitude decreases with increasing distance.
Fig 5.
Theoretical analysis of electrode size.
Figure shows SPW-Rs computed for different size electrodes (point, 100 × 100 μm2, 300 × 300 μm2, 1 × 1 mm2, 2 × 2 mm2, and 4 × 4 mm2). It indicates that high frequency oscillations and amplitudes of SPW-Rs decrease significantly with increasing surface area of the recording electrode. (b) Amplitude of sharp waves is plotted as a function of electrode area. (c) Amplitude of ripples is plotted as a function of electrode area.
Fig 6.
Microelectrode simulation was performed to observe the recorded SPW-Rs in different regions of neural network.
(a) SPW-R traces computed for a 65-electrode multi electrode array which spans a 3.5 × 3.5 mm2 total area (-1750 μm to 1750 μm). Electrodes are 100 × 100 μm2 with 500 μm inter-electrode spacing. One extra electrode is placed at the center of the network. (b) SPW-R trace calculated using a 4 × 4 mm2 electrode, comparable to clinical electrodes. (c) SPW-R traces calculated at different sites of MEA. (d) SPW-R amplitude recording as a function of distance, revealing that LFP amplitudes show a decrease as a function of electrode distance to the network.