Aims

Our brains are ready to interact with the environment the moment we are born. How can the young brain – without any prior experience – perceive and make sense of our environment? Already before our eyes open, neuronal connections have been laid out and refined to receive the first information from the outside world.

In the early fetal brain, newly generated neurons migrate to specific regions, grow axons and dendrites and then connect with each other via synapses. Molecular recognition cues guide the formation of this first coarse network. A refinement process begins, already before birth, to specify neuronal connections further not only by adding new synapses, but also through synapse elimination. Spontaneous activity, generated within the developing brain before any sensory information arrives, is thought to provide an estimate of later, sensory driven activity to fine-tune networks for processing the first sensory inputs when we are born and open our eyes. Later in life, experience driven activity further optimizes connectivity according to the environmental conditions.

Previous research demonstrated that blocking or just changing the patterns of spontaneous activity prevents normal brain development. Furthermore, spontaneous activity patterns in autistic children show clear differences compared to those in healthy children. Thus, the importance of patterned spontaneous activity for normal brain development is well established; however, the underlying mechanisms are still largely unclear.

To better understand how spontaneous activity helps wiring the healthy brain and how errors in activity-dependent wiring may cause neurodevelopmental diseases, we address the following questions.

Figure 1: H- and L-events

How is spontaneous activity generated and organized at the network level in vivo?

One hallmark of spontaneous activity in the developing brain is its correlational nature: many neurons are active at the same time. In particular, cells that will contribute to similar computational tasks in the mature brain frequently fire together already during early development.
Electrophysiological whole-cell recordings and calcium imaging of network activity with 2-photon microscopy in the visual cortex in vivo (Figure 1) allowed us to identify two types of network events.

Low-synchronicity events (‘L-events’) are characterized by the synchronous activation of subsets of cortical neurons (Siegel et al., 2012). L-events require retinal activity. We believe that they represent activity propagating from the eye to the cortex and help fine-tune cortical connectivity according to retinal inputs. During high-synchronicity events (‘H-events’), almost all cortical neurons are active at the same time. We think that these events are generated centrally, possibly within the cortex itself. The function of H-events is currently unknown, but they may be involved in processes like neuronal homeostasis or the refinement of horizontal connections.

How do individual synapses contribute to spontaneous activity in vivo?

Spontaneous activity is generated in individual neurons (e.g. pace makers) and propagates across functional neuronal networks. However, to understand how spontaneous activity refines neuronal networks, it is necessary to know also how individual synapses contribute to it. We combine patch-clamp recordings with high-resolution calcium imaging to map synaptic activity in space and time across dendrites of layer 2/3 neurons in the visual cortex in vivo (Figure 2).
Using this approach, we discovered that synaptic inputs to individual neurons are highly organized in visual cortex layer 2/3 neurons, already before eye opening (Winnubst et al. 2015). Specifically, nearby synapses often fire synchronously, whereas more distant synapses are less synchronized. This subcellular organization of inputs is called synaptic clustering.

In vivo imaging of local co-activity using 2-photon microscopy.

What are the plasticity mechanisms driven by spontaneous activity to refine connectivity?

Our previous experiments showed that synaptic clustering requires spontaneous activity to develop (Kleindienst et al. 2011), but there was no plasticity mechanism known that could explain how synaptic clustering was mediated by spontaneous activity.

Observing the patterns of spontaneous activity over time in many synapses, both in hippocampal slices and in the visual cortex in vivo, revealed a plasticity rule for synaptic clustering: synapses that are frequently co-active with their neighbors show stable activity, but locally desynchronized synapses undergo synaptic depression. Using a sophisticated closed-loop stimulation of individual synapses in hippocampal slices allowed us to demonstrate directly that local desynchronization triggers synaptic depression. We further showed that this depression is mediated by proBDNF-p75NTR signaling (Figure 3).

Together, these studies showed how spontaneous activity helps specify synaptic connections with subcellular precision. That neurons become connected so precisely has been very surprising, in particular when considering that this happens already before sensory inputs arrive. Currently, we are investigating what information is carried by those synapses that become clustered during development.

‘Out of sync: Lose your link’ mechanism

What goes wrong in neurodevelopmental diseases, such as fragile-X mental retardation or NF1 (neurofibromastosis type 1)

Besides identifying the rules of connectivity refinement in the healthy mouse brain, we also ask whether these rules may be affected in mouse models of neurodevelopmental diseases. Currently we are performing functional network imaging in Fragile X mice (a model of mental retardation and autism) and NF1 mice (a model for learning disability and ADHD) to identify changes in network activity patterns that may cause problems in network formation in these diseases.

References

• Kleindienst T, Winnubst J, Roth-Alpermann C, Bonhoeffer T, Lohmann C (2011) Activity-dependent clustering of functional synaptic inputs on developing hippocampal dendrites. Neuron 72:1012-1024.
• Siegel F, Heimel JA, Peters J, Lohmann C (2012) Peripheral and central inputs shape network dynamics in the developing visual cortex in vivo. Current Biology 22, 253-258.
• Winnubst, J., Cheyne, J.E., Niculescu, D., and Lohmann, C. (2015) Spontaneous activity drives local synaptic plasticity in vivo. Neuron 87:399-410