Developmental loss of synchronous spontaneous activity in the mouse retina is independent of visual experience

Abstract

In the immature retina, correlated spontaneous activity in the form of propagating waves is thought to be necessary for the refinement of connections between the retina and its targets. The continued presence of this activity in the mature retina would interfere with the transmission of information about the visual scene. The mechanisms responsible for the disappearance of retinal waves are not well understood, but one hypothesis is that visual experience is important. To test this hypothesis, we monitored the developmental changes in spontaneous retinal activity of both normal mice and mice reared in the dark. Using multi-electrode array recordings, we found that retinal waves in normally reared mice are present at postnatal day (P) 9 and begin to break down shortly after eye opening, around P15. By P21, waves have disappeared, and synchronous firing is comparable with that observed in the adult (6 weeks). In mice raised in the dark, we found a similar time course for the disappearance of waves. However, at P15, dark-reared retinas occasionally showed abnormally long periods of relative inactivity, not seen in controls. Apart from this quiescence, we found no striking differences between the patterns of spontaneous retinal activity from normal and dark-reared mice. We therefore suggest that visual experience is not required for the loss of synchronous spontaneous activity.

Fig. 1.

Representative spike trains recorded at different ages. At each age, spike trains from 10 simultaneously recorded cells are shown for 5 min of recording. Underneath, 15 sec expansions of the two bottom spike trains are shown.

Fig. 2.

Changes in mean firing rate during postnatal development. A, Population firing rate of all cells (P9,n = 29 cells; P11, n = 35; P15,n = 39; P21, n = 20; 6 weeks,n = 38) from one retina at each age over 180 sec in 1 sec bins. B, Histograms of the distribution of mean firing rates of individual cells. Cells from different retinas at each age have been pooled. Each mean firing rate is binned into 0.2 Hz increments from 0 to 2 Hz, except for the last bin (coloredgray), which contains all firing rates between 2 Hz and the maximum value for that age. Error bars denote 1 SEM.Arrows indicate mean firing rate for each age with the numerical value shown in parentheses.

Fig. 3.

Visualization of spontaneous activity across the retina at different ages. Each row shows 4 sec of activity (fromleft to right) at the given age. Eachframe shows the mean firing rate of each cell, averaged over 0.5 sec. Each circle represents one cell, with the radius of the circle proportional to its firing rate, subject to an upper limit of 20 Hz. The small open diamond indicates the center of mass of the active cells. Scale bar, 200 μm. See also accompanying movies (available on our website,www.jneurosci.org).

Fig. 4.

Center of mass trajectories and cross-correlations of cells at different ages. Left, Center of mass trajectory plots. Small circles indicate the approximate position of all recorded cells, with filled circles showing cells that were active within ∼1 min of recording. Three cells are numbered as they are referred to in the cross-correlation plots. Scale bar: (in P9) 100 μm (and is the same for all center of mass plots). For P9–P15, separate waves are color-coded. Center of mass was estimated every 0.5 sec.Star indicates the starting point of the trajectory. Duration of waves is as follows: P9, 3.5 sec (black), 4.5 sec (red);P11, 3 sec (black), 2.5 sec (red); P15, 17.5 sec (black), 8.5 sec (red), 14 sec (green). At 6 weeks, the trajectory of the center of mass over 30 sec of typical activity is shown; waves were no longer present. On the right are the autocorrelation and cross-correlation plots for three cells that have been marked in the center of mass plot. Horizontal axes are in seconds, and vertical axes are in hertz. The first column shows the autocorrelation of the three cells for the entire recording (50–60 min). Thesecond column shows the cross-correlation of the spike trains from a pair of cells that occurred during the period represented by the black center of mass trajectories. The third column shows the cross-correlations for the same pairs of cells for the entire recording. Above each autocorrelation and cross-correlation plot are the numbers of the cell pairs. The cyan line indicates the expectation for independent Poisson spike trains based on the entire period of recording.

Fig. 5.

Correlation indices between pairs of cells change during development. For each pair of cells recorded, we plotted their correlation index as a function of the estimated distance separating the cells. Number of cell pairs and retinas at each age are as follows:P9 (790 cell pairs from n = 2 retinas), P11 (857 pairs, n = 5),P13 (1021 pairs, n = 3),P15 (3196 pairs, n = 4),P21 (3594 pairs, n = 4), 6 weeks (6wk) (1112 pairs, n = 4). Vertical axes are plotted on a logarithmic scale. At each age, we show least-squares fits of the data to an exponential decaying function (solid line). Dashed lines surrounding the solid line indicate the 95% confidence intervals of the fit.

Fig. 6.

Effect of dark-rearing on spike trains at P15.A, Typical spike trains recorded from control and dark-reared retinas at P15 over 30 min. The mean firing rate of the population of cells is shown underneath the spike trains.B, The intervals between successive peaks in the population firing rate. The intervals are shown separately for four control retinas and five dark-reared retinas. Horizontal line indicates mean interwave interval for each retina.

Fig. 7.

Typical spike trains recorded from control and dark-reared retinas at P21 and 6 weeks over 15 min. The mean firing rate of the population of cells is shown underneath the spike trains (same as Fig. 6).

Fig. 8.

Center of mass trajectories from dark-reared retinas (DR) at different ages. Conventions are the same as in Figure 4. At P15, two waves are shown. Duration of waves was 9.5 sec (solid line) and 22 sec (dotted line). At P21 and 6 weeks (6wk), the trajectory of the center of mass over 30 sec of typical activity is shown; waves were no longer present at these ages. Scale bar, 100 μm.

Fig. 9.

The effect of dark-rearing on the correlation index at different postnatal ages. Correlation indices plotted as for Figure 5. Number of cell pairs and retinas at each age are as follows:P15 (2078 cell pairs from n = 5 retinas), P21 (1267 pairs, n = 4), 6 weeks (6wk) (3613 pairs, n = 6). At each age, we show least-squares fits of the data to an exponential decaying function (solid line). Short dashed lines surrounding the solid line indicate the 95% confidence intervals of the fit. The means of the least-squares fit from the aged-matched control data (solid lines in Fig. 5) are plotted here in long dashed lines.