How do the physiological properties of new neurons translate to a behavioral role? Are they just like mature neurons or are they unique? One idea that’s been thrown around is that their plastic period, their critical period, might endow them with an enhanced ability to associate information and contribute to memory formation. While we know that hippocampal neurons are already plastic and very capable of physiologically linking together different stimuli the big hope seems to be that maybe immature neurons are even better at this.
A related question is how fewer synapses and unique inhibitory connectivity affects their information processing capabilities. The verdict is out on whether new neurons are more or less involved in information processing than their mature counterparts. Currently, the best information we have is from studies looking at activity, measured by immediate early gene expression, in response to behavioral stimulation. The true measure of whether a neuron is involved in information processing / representation is if it spikes, i.e. fires action potentials, in response to a specific stimulus. Since new neurons have fewer synapses it’s very possible that they aren’t able to represent many different types of information, and therefore aren’t capable of associating information during memory formation. On the other hand, new neurons synapses are more plastic, perhaps making them better able to associate information even if they have fewer synapses.
And so I wasn’t surprised that the Schinder lab was the one to answer these questions at SFN. They were the ones who finally definitively showed that immature neurons are indeed more excitable than mature neurons. And on Saturday, Lucas Mongiat presented right next to me during the poster session and was kind enough, once 5 o’clock rolled around, to take a few extra minutes to tell me their new story (after traveling all night from Argentina and coming straight to his poster board from the airport!).
They first addressed the question of whether or not immature neurons are more likely to be “activated” than mature neurons. Activation was measured in retrovirally-labeled neurons in hippocampal slices by two methods: calcium imaging and spiking in response to perforant path stimulation. Input-output curves showed that immature neurons were more likely to be activated than mature neurons, regardless of the stimulus strength. Mature neurons could be made to behave like immature neurons if picrotoxin was added to the bath, blocking inhibitory connections.
So what’s going on with inhibition? At the time of spiking, inhibition and excitation were equivalent in the mature neurons, making it hard to become excited and fire an action potential. In contrast, inhibition occurred later in immature neurons, making it easier to fire in response to perforant path activity. Their next question was whether or not new neurons are more likely to associate inputs at a physiological level.
To answer this, they used two stimulating electrodes to activate different perforant path inputs onto granule neurons. Calcium responses were measured to identify activated neurons. The idea is that the two inputs will activate populations of neurons. Some neurons will be activated only by input 1 and others only by input 2. However, another population will be activated by both inputs and those neurons are the ones that are best suited for associating information as would happen during memory formation. Comparing immature (4w old) and mature (8w) adult-born cells as well as the overall population (which would be a mixed age, though mostly older than the adult-born cells) it was only the immature adult-born cells that were more likely to respond to both inputs. That the mature adult-born cells were similar to the general population suggests that adult-born cells eventually mature to become similar to perinatal-born cells.
While I’m no expert electrophysiologist, these data are exciting because they show for the first time that new neurons have the physiological potential to associate inputs better than mature neurons. Computational models and behavioral predictions needed this!
The different timecourse of inhibition makes me wonder how immature neurons handle timing of activity. For example, if input 1 is strongly activated and a subthreshold input 2 follows shortly thereafter, is input 2 potentiated and better able to activate the postsynaptic cell in the future? Does the delayed inhibition make them better at temporal summation? Does it alter the recruitment of immature neurons into cell assemblies, since this is dictated by the precise timing of neuronal firing?
DATA? MORE DATA? PLEASE?!?!