Our data indicate that this shift toward excitation may occur partially by structural changes in inhibitory neurons. A recent study in rat barrel cortex, combining

viral labeling and chronic two-photon imaging to examine structural dynamics of GAD65 positive inhibitory neurons (Marik et al., 2010), suggests that following sensory deprivation via whisker plucking there is an increase in the growth and retraction Akt activity of inhibitory neurons’ axons in the deprived and, to a slightly lesser extent, the nondeprived barrels within 2 days of deprivation (Marik et al., 2010). We did not observe axonal remodeling in our study. This may indicate differences between barrel and visual cortical plasticity. What leads to structural changes

in inhibitory neurons? For the spines on inhibitory neurons, one plausible explanation is simply that the synapses undergo long-term depression (LTD), which has been demonstrated to occur after sensory deprivation in vivo (Rittenhouse et al., 1999). Reduction of spine density has been shown to be associated with LTD in excitatory cells (Nägerl et al., 2004), but has yet to be investigated in inhibitory neurons. Mechanisms leading to the reduction of axonal boutons are less clear. Given that spine density decreases before bouton density, one possibility is that the effect is causal and a reduction of inputs to the inhibitory neuron, which presumably leads to a decrease in postsynaptic spiking, triggers a reduction of bouton density. As only see more a fraction of the boutons are eliminated, however, it is unclear which boutons would be removed and which would be spared. Some insight may come from a previous study of inhibitory of neurons in hippocampal cultures (Hartman et al., 2006): reduced activity of excitatory cells in the network caused a decrease in inhibitory synapses, but lowering activity levels of an individual inhibitory cell without altering activity in neighboring pyramidal neurons had no effect on inhibitory synapses. These results suggest that the activity

of the postsynaptic excitatory cells may be responsible for changes to inhibitory boutons. Alternatively, signaling via second messengers, such as TNF-α released from astrocytes located at synapses (Fellin, 2009 and Park and Bowers, 2010), could be involved as well. In principle, the rapid changes of inhibitory structures following focal retinal lesions could reflect the onset of functional recovery in the visual cortex, or simply be caused by reduced cortical activity levels. To distinguish between these two possibilities, we compared the effects of focal and complete retinal lesions, the latter of which are not accompanied by functional recovery in the visual cortex (Keck et al., 2008). Therefore, any structural changes following this intervention can be considered a response to a decrease in cortical activity.