In vivo Kartierung individueller Photorezeptoren und ihrer Beiträge zu rezeptiven Feldern visueller Neurone mithilfe auf adaptiver Optik basierender Mikroskopie.
Zusammenfassung der Projektergebnisse
The aim of this project was the identification of the first in vivo, cone-field maps that underlie the receptive fields (RF) of neurons in the retino-cortical pathway. The spectral sensitivity and the functional weighting of individual M and L cones were to be determined by extracellularly measuring the responses of isolated neurons in the lateral geniculate nucleus (LGN) of macaques. Stimuli with cone-sized resolution were created by minimizing optical aberrations and eye motions using adaptive optics (AO) microscopy. AO microscopy also facilitated mapping of the recorded responses to simultaneously recorded images of the retinal photoreceptor mosaic. The resulting cone-field maps were meant to address open questions on RF sizes, cone weighting, and cone opponency. Due to technical and administrative problems, data acquisition had to be paused in 2017 and was only continued recently. Most of 2018 was spent resolving the technical issues and limitations of the AO microscope and its control software. Upgrading the system has made data acquisition not only more reliable and faster, but also reduced the time spent on data analysis substantially. Data acquisition was further improved by a custom written software that allowed the online analysis of the size, retinal location, and cone composition of RFs, facilitating the targeted stimulation of single cones. The data set I collected in 2017 consists of electrophysiological recordings from 22 isolated LGN neurons of 2 adult macaques. By reverse correlating the recorded response to the binarized ‘white’ noise stimulus, I was able to extract the four-dimensional spike triggering average (STA). The RF centers extracted from the two-dimensional, spatial component of the STA were substantially smaller than RF centers measured in previous studies that could not rely on adaptive optics to minimize eye motion and optical aberrations. Correcting for transverse chromatic aberrations allowed me to map the STA to the cone mosaic. Located at 0.5 to 2.5-degree eccentricity from the fovea, the STA of most neurons encompassed 1-3 cones. Only the STA of 3 OFF cells extended over several cones. Comparing the two channels of the STA (543 and 711 nm) revealed the cone’s spectral sensitivities. The majority of cones belonged to the M type and all STAs that covered more than one cone were comprised of M or L cones only. I also measured the responses of LGN neurons to stimulation of single cones and cone pairs. The responses to the 543 nm light yielded the synaptic weighting of both M and L cones, while the 711 nm responses determined their spectral sensitivity. Interestingly, neural responses could be saturated by stimulation of single cones. My data also revealed that cones have different synaptic weightings in vivo and that STAs only represented the contributions of cones with high synaptic weight. Thus, RF sizes were underestimated when extracted from STAs. However, even with the additional cones, RF sizes were still smaller than reported in earlier studies. The responses to the 711 nm stimulus revealed that the center of some LGN neurons were formed by a combination of M and L cones. By comparing responses to stimulation of single cones and cone pairs in the RF center, I examined cone summation in LGN neurons. The cone pair responses were often smaller than the sum of the single cone responses indicating non-linear integration of inputs. However, when the rest of the RF was stimulated with binarized ‘white’ noise, responses from single cones were integrated linearly. One explanation could be that while stimulation of the RF center saturated the neuron’s response, stimulating the inhibitory surround likely moved the response back towards the linear domain of the neurons’ intensity response functions. Finally, all tested LGN neurons showed fast response adaptation. After reaching the maximum, response typically dropped to 50% within 1 frame (33 ms) before slowly declining until stimulus offset. I observed differences between ON and OFF cells with the later having a slower rise to the maximum response and a slightly faster adaptation. The results also suggested that the influence of center surround interactions on adaptation might be bigger for OFF cells than for ON cells. At this point, data currently being collected has to be analyzed and compared to the preliminary results before a comprehensive answer to the initial questions of the project can be given.
