It has been assumed that use of neurons through activity (action potential spikes) stimulates more synapses and stronger circuits in the brain and that lack of use leads to pruning or elimination of the synapse. But, is this true? While it appears that activity can lead to new and increased circuits, the mechanisms for pruning are more mysterious. New research now sheds light on how activity makes circuits, but shows that many synapses are independent of activity and are not necessarily pruned even if they are not used.
Another common view about pruning has been that different circuits compete for survival, as in natural selection. But, is this true? The new research shows great variability in how neurons respond. Many neurons and circuits function without competition. Neurons, also, can behave quite individually. This huge variability will make mapping the brain in a meaningful way much more difficult.
A major question for understanding brain function has become: where does activity determine synaptic creation and pruning?
Research in the Eye Shows Independence of Neurons and Non Competition of Synapses
A previous post, How Many Different Kinds of Neurons Are There, is based on research in the retina. The neuron types and connections in the retina have been a focus of intense study because it has a very tight, layered structure.
The retina is, also, the source of new research into how new synapses and stronger circuits are stimulated by activity. This research shows that many increases in connectivity, while related to activity, are not competitive. It, also, shows that pruning is not necessarily connected with activity of the synapse at all. This is a major shift in understanding of the brain and raises many new complications and questions.
New studies show that multiple different types of signals can converge on a single neuron. These different signals can stimulate synapse structures for many different types of circuits independent of each other. These independent synapses and circuits demonstrate individual behaviors. Actions of a single synapse, even very close to another on the same neuron, might have no impact on the other synapses nearby.
Fetal Circuit Development
There are two general ways that connections occur. Convergence is when many neurons connect to one postsynaptic neuron. Divergence occurs when one pre synaptic neuron connects to many different neurons.
Convergence is the more common outcome in the fetus. Some of these connections are pruned by inactivity and are still left with multiple connections.
The fetus sends some diverging axons to very distant places and then many of these are pruned. Some cortical neurons send long shoots to the pyramidal tract and then prune them. In the olfactory circuits, many receptors go to several glomeruli and then later leave only one. This same thing happens in the eye where bipolar cells first connect with ganglion cells, but later end up with two types of amacrine cells.
Therefore, refinement of circuits and activity does help establish the adult pattern in the fetus. The previous theory has been that this occurs through competition. But, in fact, there are many that are non competitive.
Non-Competition in Forming Connections
Axons from the eye connecting to layer 4 of the visual cortex are arranged very specifically into alternating columns of equal size related to each eye. In early research, it was noted that if one eye is closed, the relevant column shrinks and the good eye region enlarges. This encouraged the idea that alterations in signaling create differential circuits.
Recent research shows subtle changes in signaling intensity provides different kinds of circuits. The timing of the spikes, not just the intensity, appears to be a relevant factor in creating synapses and circuits.
Of two major layers in the retina, one responds to light of different levels, and the other sends the opposite signal. Both layers signal to the same ganglion cells. It is important to note that changing one signal doesn’t affect the others, even on the same ganglion cell.
They do not compete, but rather cooperate in the same synapse. More synapses are formed with more activity for the type involved. The other connections to the same ganglion cell don’t alter their synapses. This is an example where more signals create more synapses, but the process doesn’t dominate or reduce the other connections. There is no competition.
There are other structures in the eye where activity modifies synapses but not through competition. One type of retinal cell responds to horizontal movement and has inputs special for ultraviolet light signals and others for blue light. If the ultraviolet cells are interfered with, there are more blue cell connections. But, there was no competition because the ultraviolet connections stayed the same.
These examples demonstrate lack of competition in circuits that do respond to activity.
Uniform Signaling and Different Results
There are many examples of different neuronal types having different types of increased connectivity even though there is a general increase of activity for all of them.
In one region of the retina, there are seven types of a particular bipolar neuron (see post on How Many Kinds of Neurons Are There). If all of the cell types have the same amount of electrical activity, some of the neurons end up with more increased connectivity than others. When there is a decrease of sensory input for all of them, some types increase more than others.
The mechanism where increased activity creates more circuits appears to be quite variable. Sometimes they stimulate more circuits and sometimes not.
One view of these phenomena is that competitive pruning only occurs in the fetal brain when initially all of the axons and their connections are mingled together. In many circuits, activity can create new connections, while those that are not active wait for their turn and then create a different circuit.
Another situation where competition could occur is when distant neurons all connect to the same neuron. But, in fact, competition doesn’t occur when neurons send axons to the same cell, but to completely different dendrite sections. If one synapse has more activity, it can create more connectivity without altering the other neurons.
In very structured layers such as in the retina, there is less competition.
Tiling Produces Non Competition
Another non-competitive situation is “tiling,”—multiple connections where each axon arbor has no overlap with the others. This type of circuit has multiple axon and dendrite arbors, but they are not redundant and provide little overlap; this provides maximum coverage and efficiency for the region. In other words, with tiling, they do not compete.
However, if some cells are damaged, those that are not damaged create new abnormal connections in the all of the territory and don’t continue tiling. These abnormal axon connections are called sprouting.
Other research shows that even when there is great overlap of connections, they don’t compete. Two types of bipolar cells overlap greatly and all of their synapses are intermingled with a particular type of ganglion cell. These individual neurons can produce a circuit without influencing the other overlapping cells.
In the retina, disruption can cause increased dendrites as well as some retraction of axons. Some produce sprouting with branches that are very different from the previous pattern.
When the eye first opens at birth, this same type of sprouting occurs. Photoreceptors and their postsynaptic cells send out abnormal sprouts. In abnormal situations of regeneration the findings can be extremely variable.
When Is There Competition?
There is evidence that neurons in some regions have limited energy and therefore choose one particular circuit at the expense of others. Choosing one definite circuit would then limit connections to other circuits. This occurs in the cerebellum and the neuromuscular junctions of muscles. Both involve competition.
There are other regions, such as the retina, where the resources are not limited. In the retina, a neuron can greatly increase its connectivity to a very high density and this doesn’t impede other connections. There is no competition.
Pruning Unrelated to Activity and Competition
While some circuits do prune synapses that are not used, many areas, including the retina behave differently. In the retina, pruning can occur even with continued activity. One example of pruning despite activity is the connection between specific rod bipolar cells and ganglion cells. These synapses are pruned whether there is a lot of activity or no activity. By adulthood, these connections are all gone.
Another new finding in the retina is when there is a lot of activity and the rate of making new synapses increases. This is, also, demonstrated in the hippocampus. But, in these situations, the rate of pruning is not affected at all.
The rate of making new synapses between bipolar and ganglion cells in the retina is lower when there is less activity. With more activity, it is increased. But, still there is no alteration in the pruning of synapses at the ganglion cells.
Specialized Non Competitive Synapse Structures – Ribbon Release and PSD95
The exact mechanisms for the increase of synapses are not clear. Dendrites behave like an amoeba and produce filopodia in different directions. These protrusions become dendritic spines and are able to attract axons and then connect with them. The mechanism for increased synapses has been thought to involve neurotransmitters producing bulges from the dendrites and axons.
But, in the retina there are much more complex mechanisms making synapses that use very unique complex structures—the ribbon release on the pre synaptic neuron and PSD95 on the past synaptic neuron. These are non-competitive and produce very unique variable results. It provides very rapid neurotransmitter release needed for vision.
In the retina special structures allow a graded release of the neurotransmitters and reside in a presynaptic region called the active zone. The ribbon release sites are in the active zone. They are perpendicular to the membrane and hold a large number of vesicles for rapid neurotransmission.
On the postsynaptic cells, unique structures in the postsynaptic density—PSD95—meet the ribbon release sites. PSD95 clusters are different in each region. They consist of protein enzyme complexes that are part of the post synaptic density and sit near the membrane. PSD95 is a scaffolding structure that brings together many receptors, ion channels and other signaling proteins. It is linked to neuroligin and is very important in neuroplasticity by maintaining the synapse structure during changes in neuroplasticity.
The ribbon release in the pre synaptic neuron rapidly releases and re absorbs multiple vesicles at the same time. Release is triggered by highly regulated graded calcium fluctuations. It is a very accurate rapid process which is useful for complex sensory systems such as vision and hearing. It appears in photoreceptor cells and bipolar cells of the eye, and cochlear hair cells and vestibular receptors of the ear.
With ribbon release/PSD95, neurotransmitter activity adds new synapses at already stable parts of the synapse. This is an important example of synapses increasing and being pruned that is variable and not competitive.
The Effects of Neuron Activity Varies with Age
The effects of electrical activity on neurons varies at different ages—in the fetus, children, and adults. At these different ages, there are different actions related to forming and pruning of synapses. Even in the same regions, activity causes different actions in forming synapses, and pruning.
Before the critical period in childhood, if one eye is blocked so that the activity is only in the other eye, dendrite spines are created less frequently, but there is no pruning in the visual cortex pyramidal neurons.
Early transmission of bipolar and ganglion cells affects creation of new synapses, but not pruning. In the critical period where new structures form for vision, however, lack of activity leads to increases in spines and more pruning.
Both in the fetus and adult, some synapses can be altered by increased and decreased activity. Changes include producing different circuits, more synapses and structural and functional alterations of the synapses. In some circuits decreased activity changes the rates subtly or increases the sizes of the spikes.
Different regions and even different individual neurons have very different responses to decreased activity.
In the retinal bipolar cells there are changes in both the pre and post synaptic neurons with decreased activity. Postsynaptic densities actually increase their size (as if searching for more activity). More ribbons sometimes occur.
In the retina, the synapses between bipolar cells and ganglion cells change individually based on activity and often in the opposite direction than would be expected. There is no balancing effect and no effects on nearby neurons and synapses.
Another place this happens is in the visual cortex layer 2, 3 and 4 where decreased activity causes increase in the amplitude of the spikes on the postsynaptic cell. In receptor 6 of cone bipolar cells the opposite occurs.
Differences On The Same Neuron
But, even more striking are differences in varied synapses on one neuron. Basket interneurons in Layer 4 pyramidal neurons decrease the size of the action spike, but other connections on the same neuron increase.
Separate connections on the same neuron can have quite different responses.
Are These Differences Determined By Being Part of a Larger Circuit?
Most often inhibition and stimulation are determined together. But, there are specific cells such as rod bipolar cell A17 that do not work together; each is done individually. On this one neuron, the excitation and inhibition regions are right next to each other but they respond completely independently.
All of this shows that the regulation of rates of firing are different in varied synapses and individual neurons. In the retina, individual synapses change individually and uniquely.
Does Activity Determine Synaptic Creation and Pruning
With increased ability to observe individual neurons in animals, research is showing very unusual variety in the responses of neuron types and even individual neurons of the same type. This great variation in the behavior of individual neurons will greatly complicate mapping of the brain.
In searching for rules, what is found is great variation in cell types, in synapse types (inhibitory and excitatory), single or multiple ribbon-PSD95 sites and different neurotransmitter types. The variation appears even in the individual neurons and individual synapses on one postsynaptic neuron.
Another recent finding is that activity alters some prominent neurons in a system in the same way, but not others in that circuit or its neighbors.
All of the rules break down when there is damage to a nerve.
As in many other parts of neuroscience, as more research is done, more complex and variable patterns are found. Where is the direction for all of this individual neuron variability?
The increased complexity and seemingly individualistic behavior of neurons are more consistent with the view that there is no brain center, but, rather, mind interacts with each neuron and each molecule in the brain.