Contrasting extracellular signatures associated with active dendritic chemical synapses vs. gap junctions receiving synchronous inputs.

A. Top, 3D electrode setup representing 7×7×7 (343 in total) electrode array spanning the basal dendrites of a CA1 pyramidal neuron morphology. Although the entire morphology was used for simulations spanning the apical and basal dendrites, the depiction here is restricted to the basal dendrites to emphasize electrode locations. Bottom, field potential traces from electrodes at proximal (15 traces representing different locations within 50–100 μm) and distal (21 traces representing different locations within 190–300 μm) locations along the somato-basal axis, when dendrites were active (left) or passive (right). Black traces depict the respective average trace across all distal or proximal locations. The 245 chemical synapses (Nsyn = 245) which were randomly dispersed across the basal dendrites received synchronous inputs. B. Same as A, but for external inputs arriving through gap junctions. The number of gap junctions Njun = 217. C. Amplitudes of negative deflection of field potentials for all 343 electrodes, plotted as functions of radial distance of the electrode from the soma, for active and passive dendritic models receiving synchronous inputs through chemical synapses (Nsyn = 245). Inset shows plot of median field potential amplitude values as a function of distance for both active and passive dendritic configurations. D. Same as C, but amplitudes of positive deflections in extracellular potentials associated with inputs arriving through gap junctions. The number of gap junctions Njun = 217. Comparison of active vs. passive dendritic configurations in (C–D): * p<0.05, **p<0.01, ***p<0.001, Wilcoxon rank-sum test. E. Extracellular electrodes were placed across the entire span of the neuron (active dendrites with no sodium) instead of being confined to the basal dendritic span (panels A–D), with all parameters set identical to panels A–D. A flip in the sign of the extracellular potentials may be noted for synchronous stimulation with chemical synapses (Left), but not with stimulation with gap junctions (Right).

Differential polarity of field potentials associated with synchronous inputs through chemical synapses vs. dendro-dendritic gap junctions on active dendrites.

A. Extracellular potentials from electrodes at proximal (15 colored traces represent different locations within 50–100 μm) and distal (21 colored traces represent different locations within 190–300 μm) locations along the somato-basal axis. Shown are traces for default (where all components were present), no sodium, no leak, and no sodium or leak scenarios for active dendritic structures. Black traces in each scenario depict the respective average trace across all distal or proximal locations. The 245 chemical synapses (Nsyn = 245) which were randomly dispersed across the basal dendrites received synchronous inputs. B. Zoomed example trace (from a proximal electrode at 54 µm from soma) showing the impact of leak channels in shaping the extracellular potentials associated with active dendritic structures with (Default) and without leak channels (No leak). C–E. Mean and SEM of the amplitudes of negative deflection (C), positive deflection (D), and the total peak to peak amplitude (E) of the extracellular potentials, plotted as functions of radial distance of electrode location, for default, no sodium, no leak, and no sodium or leak scenarios for active dendritic structures. F–J. Same as panels A–E, but for active dendritic structures receiving synchronous inputs through dendro-dendritic gap junctions (Njun = 99).

Ionic basis of the differential contributions of transmembrane currents to field potentials associated with synchronous inputs arriving through chemical synapses vs. gap junctions on active dendrites.

A. Mean and SEM of peak membrane voltages (top) and peak synaptic currents (bottom) from across somato-basal locations recorded intracellularly for all 4 model configurations. The 4 different model configurations shown are the default active model, and the active models where sodium channels, leak channels, or both sodium and leak channels were absent. It may be noted that there were no action potentials or dendritic spikes when there were no sodium channels in the models. The dependence of synaptic current on the membrane potential, acting as the driving force, may also be noted. B. Mean and SEM of peak values of transmembrane sodium, calcium (T-type, L-type, R-type, and N-type), HCN, leak, capacitive, and potassium (A-type, delayed rectifier, and M-type) currents for different active models receiving synchronous inputs through chemical synapses, plotted as functions of radial distance from soma for all 4 model configurations. C–D. Same as panels A–B, but with different configurations of active models receiving synchronous inputs through gap junctions. There are no synaptic currents plotted here as there are no transmembrane synaptic currents associated with gap junctions.

Differential spatiotemporal structure of field potentials associated with active dendrites receiving low-frequency random inputs through chemical synapses vs. gap junctions.

A. Distance-wise LFP responses to low-frequency random inputs (LFRI) impinging on active dendrites through chemical synapses. Rows 1–3: LFP data from electrodes located at a distal (∼152 µm; Row 1), intermediate (∼97 µm; Row 2), and proximal (∼55 µm) locations with reference to their radial distance from the soma. Column 1: time-domain signal. Column 2: Fourier transform of the signal shown in Column 1. Column 3: spectrogram of the signal shown in Column 1 computed using wavelet transform. B. Same as panel (A) but for active model lacking sodium conductance. C–D. Same as panels (A–B), except low-frequency random inputs impinged onto active dendrites through gap junctions.

Dominance of specific oscillatory bands in field potentials depended on whether inputs onto active dendrites were received through chemical synapses or gap junctions.

A. Example LFP responses to rhythmic inputs at different (1–64 Hz) frequencies and their spectral signatures, shown for simulations performed with active or passive basal dendrites receiving rhythmic inputs through chemical synapses. Each row shows the filtered LFP signal at an electrode placed ∼97 µm from the soma, the Fourier power spectrum, and the wavelet spectrogram for the LFP signal. Different rows depict different input frequency values for the rhythmic input (1 Hz, 4 Hz, 16 Hz, and 64 Hz). B. Same as (A) but for the rhythmic inputs impinging on basal dendrites at different frequencies through gap junctions. All simulations depicted here were performed in the absence of sodium channels to avoid spiking.

Distance-dependence of spectral power in specific bands of field potentials associated with active dendrites receiving rhythmic inputs through chemical synapses or gap junctions.

A. Maximum power in LFP responses associated with rhythmic inputs at different (1–128 Hz) frequencies, shown for simulations performed with active or passive basal dendrites receiving these rhythmic inputs through chemical synapses. All electrodes at specific radial distances are depicted for each scenario. The frequency of the rhythmic input is highlighted in each panel. B. Same as (A) but for the rhythmic inputs impinging on basal dendrites at different frequencies through gap junctions. Across all plots, lines connect the respective median values (represented by black stars). All simulations depicted here were performed in the absence of sodium channels to avoid spiking. * p<0.05, **p<0.01, ***p<0.001 (Wilcoxon rank-sum test)

Transmembrane currents driven by voltage responses mediate the differential emphasis of specific oscillatory bands in field potentials associated with chemical synapses vs. gap junctions on active dendrites.

A. Row 1: Distance-dependent maximal power of local field potentials recorded at different electrodes (shown as mean and SEM) associated with neuronal response to rhythmic inputs at different frequencies impinging on the active basal dendritic model through chemical synapses. Row 2: Fourier power spectra for all field potentials at different frequencies of the rhythmic inputs. Each trace for a given frequency represents different electrodes. Row 3: Fourier spectra of the filtered total transmembrane current for each frequency of rhythmic inputs, from each basal dendritic compartment. B. Same as panel A, but for simulations performed with passive dendrites. C–D. Same as (A–B), but with rhythmic inputs coming through gap junctions. All simulations depicted here were performed in the absence of sodium channels to avoid spiking.

Differential phase relationship between local field potentials and spikes associated with rhythmic inputs through chemical synapses vs. gap junctions on active dendrites.

A. Example LFP traces (color coded based on whether they are associated with chemical synapses in dark pink or gap junctions in blue) and simultaneously recorded intracellular somatic voltage traces (Black) for neurons receiving rhythmic inputs through chemical synapses (top) or gap junctions (bottom). Shown are traces with default model configuration where all channels were intact (left), traces where the intracellular traces were filtered to the respective band (middle), and extracellular/intracellular traces obtained in the absence of sodium channels (right). B. Left, Spike phase with reference to local field potentials for each spike (lighter circles) for oscillatory inputs at different frequencies impinging on active dendrites through chemical synapses vs. gap junctions. Dark-colored circles represent the median values at each frequency for the respective group. Right, Polar version of the plot showing median of spike-LFP phases over five trials for rhythmic inputs at different frequencies through chemical synapses and gap junctions. The different frequencies are represented along the concentric circles and the corresponding spike-LFP phase values are plotted along the angular axis. C. Polar plot with median of phases obtained from cross-correlation of filtered intracellular potential and corresponding LFP traces at respective frequencies over all trials. These plots were derived from the same intracellular traces as in panel A but represent phase differences between extracellular traces and the entire filtered intracellular voltage trace. D. Polar plot of phases obtained from cross-correlation between intracellular potential (without sodium conductance) and corresponding LFP traces when oscillatory inputs were presented through chemical synapses or gap junctions.