, 2001 and Safran et al., 2007). This large time constant introduces a long-lasting ringing in response to the onset of motion (A.B., unpublished data), incompatible with the observed cellular responses (Egelhaaf and Borst, 1989 and Reisenman et al., 2003). A further conflict arises in this model’s prediction of negative responses to ON-OFF and OFF-ON pulses, which are clearly absent in the experimental data (Eichner et al., 2011). Nevertheless, while we feel that there is evidence arguing against the six-detector model
Rapamycin concentration with mixed channels, definitive clarification of the discrepancies will require further direct investigation. Taken together with the evidence of the pathways leading from L1 to T4 and from L2 to T5 cells, respectively, our current view is that in the fruit fly, two separate motion detection systems operate in parallel, one analyzing the movement of light increments and the other one the movement of light decrements (Figure 4H). While the exact nature and role of the participating neurons is still unclear, the splitting of the positive- and negative-going
brightness signal into two channels, one for signals of the positive, the other for signals of the negative sign, has interesting consequences for the multiplication as postulated in the Reichardt detector. Without splitting, the output of such a putative multiplication neuron would need to increase in a supralinear way when both input signals go positive
as well as when they go negative. Such a mechanism is difficult to realize. 5-Fluoracil chemical structure However, splitting the input into separate channels leads to positive signals only, and while a number of biophysically plausible mechanisms have been proposed to do that (Torre and Poggio, 1978, Srinivasan and Bernard, 1976, Gabbiani et al., 2002, Hausselt et al., 2007 and Enciso et al., 2010), the exact mechanism active within these neurons presynaptic to the fly LPTCs remains to be determined. In a similar way, the biophysics underlying the temporal filtering as postulated by the Reichardt detector already represents another challenge for future research. Since the majority of studies focused on ON/OFF DS cells and their circuitry, we will concentrate on those, while only briefly touching upon other types (see Mechanisms in Other Types of Retinal DS Ganglion Cells). The original Barlow-Levick model (Barlow et al., 1964; reviewed in Masland, 2004) proposed that DS ganglion cells receive delayed and/or long lasting inhibition preferentially from interneurons displaced to the null side of their dendritic field. Note that the term “null/preferred side” refer to the positions from which the null/preferred direction stimulus enters the ganglion cell’s dendritic field. This inhibition would be triggered by a stimulus moving in the null direction toward the cell’s receptive field center and would cancel out any excitation caused by the stimulus when it eventually enters the center.