Distinct superficial and deep laminar domains of activity in the visual cortex during rest and stimulation

Abstract

Spatial patterns of spontaneous neural activity at rest have previously been associated with specific networks in the brain, including those pertaining to the functional architecture of the primary visual cortex (V1). However, despite the prominent anatomical differences between cortical layers, little is known about the laminar pattern of spontaneous activity in V1. We address this topic by investigating the amplitude and coherence of ongoing local field potential (LFP) signals measured from different layers in V1 of macaque monkeys during rest and upon presentation of a visual stimulus. We used a linear microelectrode array to measure LFP signals at multiple, evenly spaced positions throughout the cortical thickness. Analyzing both the mean LFP amplitudes and between-contact LFP coherences, we identified two distinct zones of activity, roughly corresponding to superficial and deep layers, divided by a sharp transition near the bottom of layer 4. The LFP signals within each laminar zone were highly coherent, whereas those between zones were not. This functional compartmentalization was found not only during rest, but also when the receptive field was stimulated during a visual task. These results demonstrate the existence of distinct superficial and deep functional domains of coherent LFP activity in V1 that may reflect the intrinsic interplay of V1 microcircuitry with cortical and subcortical targets, respectively.

Keywords: LFP; V1; coherence; gamma; laminae; layers; resting state; visual perception.

Figure 1

Schematic representation of laminar LFP recordings from a linear multicontact electrode array in the primary visual cortex. For illustrative purposes, the array is depicted overlying a Nissl-stained histological slice, with labels showing individual layers and the corresponding compartments (i.e., supragranular, granular, and infragranular) used in the study. On the right is a sample of extracellular LFP data collected in one experiment. Each trace corresponds to the voltage measured simultaneously as a function of time (see scale bar). In this example, the top contacts span an area reaching from outside the brain (top) to the white matter (bottom). Note the spatial non-uniformity of gamma frequency signals, superimposed on the signals, and restricted to the upper cortical layers.

Figure 2

Current source density (CSD) and spectral profile of LFP magnitude as a function of laminar position for one session. (A) CSD profile in response to a flickering screen collected during an example session (monkey CB35). The center of the initial current sink, thought to correspond primarily to thalamic input in layer 4Cα, was used as a point of alignment throughout the study. Dotted lines ±250 μm correspond roughly to the extent of layer 4. (B) Spectral laminar profile during rest. On the left is a plot depicting mean LFP magnitude for a representative 20-min period following the CSD experiment with the monkey at rest. Signal magnitude is color-coded on a log scale, and plotted as a function of frequency and cortical depth. On the right is the same data, normalized by the mean spectrum across all layers, and expressed as percent deviation from this mean. Note the elevated LFP magnitude in the granular and deep supragranular layers (arrow). This feature proved highly consistent across sessions (see Figure 3A).

Figure 3

Intersession average following CSD-based realignment. (A) Laminar differences as a function of frequency over all sessions. (B) The mean LFP magnitude during resting state over all sessions following alignment, expressed as a function of frequency. The dashed white line represents the “zero point,” with the dotted black lines showing rough boundaries of layer 4. The elevated high-frequency activity in the middle and upper layers is clearly visible. The colored bars indicate the frequency range used to compute mean LFP magnitude in (C). (C) Laminar distribution of LFP amplitude in gamma and sub-gamma-ranges during rest. Note that the mean gamma-range amplitude is highest in the middle and upper layers.

Figure 4

(A) Inter-contact coherence for reference contacts taken from two different laminar compartments. Magnitude-squared coherence in the gamma-range (30–100 Hz) is plotted between each reference electrode and all other electrodes in the array (including contacts located in the white matter at the bottom of the array and outside the brain at the top, respectively). The electrode from the granular zone (red) showed strong correlation with granular and supragranular sites, but weaker coherence with infragranular sites. In contrast, the infragranular contact (green) showed strong coherence with the infragranular contacts, but very low coherence with other sites. Note that the two electrodes chosen for this example (E₀, red and green) are separated by only 200 μm. (B) Laminar pattern of inter-contact coherence for 10 different E₀ contact positions, shown in the same format as in (A). The E₀ positions in the infragranular layers elicit a pattern of high coherence only in those layers, suggesting that the gamma-range activity in those layers is highly synchronous, but asynchronous to that in other layers. Conversely, E₀ positions in the supragranular layers are coherent only with signals measured in supragranular contacts. A single contact lying just below the zero point appears to be a transition between supra- and infragranular coherence.

Figure 5

Pair-wise coherence of all sessions in the gamma-range. (A) Average CSD to flashing screen following alignment to initial sink. (B) Mean gamma-range coherence, computed between all pairs of laminar positions over all sessions, during rest (see Figure 7 for effects of visual stimulation). The red regions reveal the high inter-compartmental coherence, with the adjacent blue regions revealing the lack of coherence between compartments.

Figure 6

Laminar coherence as a function of frequency (n = 13, session; both monkeys). Inter-compartmental coherence for the classic EEG frequency bands (delta = 1–4 Hz, theta = 5–8 Hz, alpha = 9–14 Hz, beta = 15–30 Hz, low gamma = 30–50 HZ, high gamma = 50–100 Hz) is plotted individually using the same format as Figure 7. Note that despite differences in the overall coherence, the basic pattern between upper and lower layers remained.

Figure 7

Interlaminar coherence during visual stimulation (n = 13 sessions, both monkeys). All conventions are the same as in Figure 5B. (A) Coherence pattern during fixation before stimulus onset (−300 to 0 ms before stimulus onset). (B) Coherence pattern during sustained presentation of a luminance stimulus onto the receptive field (900–1200 ms after stimulus onset). (C) Coherence pattern following the removal of the stimulus (600–900 after stimulus offset). Note that despite differences in the overall coherence level compared to the resting condition (Figure 5), the spatial pattern between upper and lower layers was similar.

Figure 8

Pair-wise coherence of the slow fluctuations in gamma power computed for all sessions (lasting 20 min each). Data presented in same format as Figure 7, but now pertaining to 0.01–0.1 Hz fluctuations in the magnitude of the gamma-range LFP activity. Note these fluctuations show moderate background coherence (i.e., the blue in the plot is roughly 0.5). However, as with the voltage coherence shown above, the power coherence is highest within the same laminar compartment.