Effects of Neuronic Shutter Observed in the EEG Alpha Rhythm

Abstract

The posterior alpha (α) rhythm, seen in human electroencephalogram (EEG), is posited to originate from cycling inhibitory/excitatory states of visual relay cells in the thalamus. These cycling states are thought to lead to oscillating visual sensitivity levels termed the "neuronic shutter effect." If true, perceptual performance should be predictable by observed α phase (of cycling inhibitory/excitatory states) relative to the timeline of afferentiation onto the visual cortex. Here, we tested this hypothesis by presenting contrast changes at near perceptual threshold intensity through closed eyelids to 20 participants (balanced for gender) during times of spontaneous α oscillations. To more accurately and rigorously test the shutter hypothesis than ever before, α rhythm phase and amplitude were calculated relative to each individual's retina-to-primary visual cortex (V1) conduction delay, estimated from the individual's C1 visual-evoked potential (VEP) latency. Our results show that stimulus observation rates (ORs) are greater at a trough than a peak of the posterior α rhythm when phase is measured at the individual's conduction delay relative to stimulus onset. Specifically, the optimal phase for stimulus observation was found to be 272.41°, where ORs are 20.96% greater than the opposing phase of 92.41°. The perception-phase relationship is modulated by α rhythm amplitude and is not observed at lower amplitude oscillations. Collectively, these results provide support to the "neuronic shutter" hypothesis and demonstrate a phase and timing relationship consistent with the theory that cycling excitability in the thalamic relay cells underly posterior α oscillations.

Keywords: EEG; alpha rhythm; neuronic shutter; visual conduction delay; visual-evoked potential.

Figure 1.

A, According to the cruciform model of the calcarine fissure in the V1 cortical area, the lower portion of the fissure will respond to upper visual field stimuli and vice versa for the upper portion of the fissure. For this reason, and as the upper and lower portions of the fissure contain opposing dipoles, upper and lower visual field stimuli will create waveforms of opposing polarity in the EEG. B, Upper and lower visual field stimuli used for C1 VEP task are centered 3° below the fixation point to account for the overrepresentation of the lower visual field in the calcerine fissure as shown in A. C, C1 VEP from a single participant in response to the upper and lower stimuli, the difference was measured as the lower minus the upper VEP. The peak latency of the difference wave (indicated as +) was used as the estimate of t2.

Figure 2.

Example of the staircase method used to determine the threshold intensity in a single participant. Each time the participant observed the stimulus, the intensity in the next trial was decreased. If the participant did not observe the stimulus in one trial, the intensity was increased in the next trial. Threshold intensity was determined as the average intensity of all trials (trials 6–40 in this example) following the third reversal, or corner, in the trial-intensity trace (trial 5 in this example). The dashed threshold intensity line in this figure spans the trials over which the intensities were averaged.

Figure 3.

Each of the 20 participants’ α peak-to-peak amplitude distribution for all trials used in the analysis at t2.

Figure 4.

C1 VEP component averaged across all participants in response to upper and lower visual field stimuli. The difference waveform is the lower waveform subtracted from the upper waveform. The mean and range of each participant’s estimated retina-to-V1 conduction delay (t2) is indicated and was calculated as the peak latency of the C1 component in each individual’s difference waveform.

Figure 5.

Mean change in OR (ΔOR) for each amplitude and phase condition relative to mean OR across all conditions when analysis was conducted at t2 (error bars indicate standard error). Tukey’s HSD tests were performed between phase levels within amplitude conditions with significance indicated by *. Bonferroni corrected t tests compared each mean to zero with significance indicated by #; 90° = peak; *,#p ≤ 0.05; **,##p ≤ 0.01; ***,###p ≤ 0.001; ****,####p ≤ 0.0001.

Figure 6.

ΔOR as a function of phase for each participant in the high-amplitude and low-amplitude conditions with radial axis extending to negative values for analysis conducted at t2. The preferred phase is indicated on the circumference, calculated as the direction toward the circle’s center of mass; 90° = peak.

Figure 7.

Analysis conducted at t2. A, C, ΔOR as a function of phase, with radial axis extending to negative values. B, D, Individual participants’ preferred phase angles; their magnitude is proportional their individual PPE. The width and height of the shaded rectangles indicate, respectively, the x and y 95% CI in cartesian space. In all panels, the group mean preferred phase is indicated in black on the circumference of the diagrams, with the shaded region indicating estimated 95% CIs for each amplitude condition; 90° = peak.

Figure 8.

Change in α amplitude, relative to a −300 to −100 ms baseline interval (A), and α phase coherence for observed, missed, and all stimuli together (B). The arrow indicates a peak in the missed trial phase coherence reflecting the time point were α phase best predicts that a trial will be unobserved by the participant. There was no indication of stimulus-induced effects at the t2 time point where phase and amplitude measures were hypothesized to be of mechanistic relevance.

Figure 9.

Analysis conducted at t = −100 ms. A, C, ΔOR as a function of phase, with radial axis extending to negative values. B, D, Individual participants’ preferred phase angles; their magnitude is proportional their individual PPE. The width and height of the shaded rectangles indicate, respectively, the x and y 95% CI in cartesian space. In all panels, the group mean preferred phase is indicated in black on the circumference of the diagrams, with the shaded region indicating estimated 95% CIs for each amplitude condition; 90° = peak.

Figure 10.

Mean change in OR (ΔOR) for each amplitude and phase condition relative to mean OR across all conditions when analysis was conducted at t = −100 ms (error bars indicate standard error). Tukey’s HSD tests were performed between phase levels within amplitude conditions with significance indicated by *. Bonferroni corrected did not find any of the means to significantly differ from zero; 90° = peak; *p ≤ 0.05.

Figure 11.

Visual pathway from retina to V1. The LGN/RN negative feedback loop provides a mechanism of cycling LGN excitability state. This LGN/RN network is known to give rise to sleep spindles. The same, or a similar, mechanism is expected to give rise the posterior α rhythm.