Putative retinal and cortical potentials from melanopsin sensitive ganglion cells

To the Editor:

Intrinsically photosensitive retinal ganglion cells (ipRGCs) absorb short wavelength light via the photopigment melanopsin [1]. ipRGCs are critical for photo-entrainment, circadian rhythms, pupillary responses, alertness, and cognition, which persist in blindness or the absence of visual perception [2]. Moreover, Dacey et al. showed that “giant” ipRGSs send signals to the LGN and visual cortex substantiating their role in visual perception [3]. These giant ipRGCs show a unique opponent S-cone off response wherein S cones inhibit ipRGC firing rate while L and M cones enhance it [3]. High contrast stimuli and silent substitution revealed long-latency ERGs from ipRGCs which can be decreased in glaucoma [4]. However, high contrasts can inadvertently stimulate other photoreceptors lessening selectivity of pathway activation. We used selective chromatic adaptation, similar to SWAP perimetry, wherein a blue stimulus, effective for S cones and ipRGCs, was presented on an amber rod and LM cone saturating background to isolate S cone and putative responses from ipRGCs over a 4 log unit luminance range.

Fourteen healthy adults (mean 24.2 ± 1.6 years, 11 females, 3 males) participated in our IRB approved protocol after providing informed consent. A calibrated Ganzfeld (Diagnosys, LLC) was used to fully illuminate the retina with a 200 ms blue flash (448 nm) presented on a constant bright amber background (590 nm, 560 cd/m²). Simultaneous flash ERGs (DTL fiber electrode at lower limbus of right eye referenced to right earlobe) and VEPs (active electrode 1 cm above inion, referenced to forehead, common ground: left temple) were recorded after 30 s of adaptation to the amber background. Signals from the right eye (left eye occluded) were band-pass filtered (0.612–10 Hz) to isolate low frequency ipRGC responses [4] and recorded in 1000 ms epochs as the average response to 30 artifact free blue flashes (1 flash/2 s). Separate recordings were obtained at four blue intensities spanning 4 log units (16.7–0.0167 cd/m²). Digital values from all subjects (µV vs. ms) were averaged to compute mean ERG and VEP waveforms. The absence of rod and LM cone ERGs in waveforms and non-recordable ERGs to the ISCEV scotopic flash (0.01 cd/m²/s) on the amber background confirmed S cone and putative ipRGC response isolation.

Figure 1a shows mean ERGs at the highest luminance and at 100× lower luminance (Fig. 1b). The early small amplitude S cone ERG is followed by a wave of negativity presumed to derive from S cone inhibitory input to giant ipRGCs [3]. A positive response occurs within this negativity representing intrinsic ipRGC light onset or possibly S cone offset responses. The subsequent positivity is likely prolonged spiking by ipRGCs [3]. Figure 1c, d shows flash VEPs with the first positive peak (P1) attributable to ganglion cell input since P1 is selectively decreased in glaucoma and vascular disease [5]. Figure 2 shows mean VEP P1amplitude at four luminances plotted against corresponding ERG negative to positive peaks to index ipRGC activity. The exponential increase (r² = 0.91) in putative retinal and cortical responses from ipRGCs exemplifies the potential of this novel approach for further revealing their unique functions.

Fig. 1: ERG and VEP Waveforms.

a Shows the mean digitized ERG from 14 subjects. The S cone ERG followed by putative components of the ipRGC response are labeled. b Shows the ERG at a 100× lower blue flash luminance. c Shows the flash VEP at the highest blue luminance and 1d the VEP at 100× lower luminance.

Fig. 2: VEP vs. ERG Amplitudes.

The mean VEP P1 amplitude is plotted against the mean ERG negative trough to peak amplitude across a 4 log unit range of luminances.