Summary
This is a one-page summary of my attempt at a coherent explanation
of the working of the human brain. As described in the
introduction, it is based on seven hierarchical
levels of description of functionality within the brain.
The table below is a summary of the levels; more details on each level can be seen lower down this page.
A simple summary intended for children or
for those learning English is also available.
The structure of this website is also hierarchical,
so clicking on any link that goes to another page will give many more details, including examples and links to relevant articles,
books and scientific research; I have deliberately not put any references or external links on this page.
The seven levels
| Summary of the levels of description
of the workings of the human brain, starting with the “lowest” level
|
| 1
|
The physical constituents of the brain and their individual capabilities: neurons,
their synapse connections, and glia.
The way that signals are created in, and passed between, neurons involves the movement of ions.
Neurotransmitters are the chemicals that relay these signals,
and their use over a wider area is called neuromodulation.
|
| 2
|
Memory-enhanced coincidence detection and lateral inhibition is a function
that emerges from the most basic capabilities of neurons and synapses.
Coincidence detection is best explained by considering a simple model neuron.
|
2-5
Levels 2 through to 5 are different
levels of description of the
processing of data, from
both external and internal senses,
and from within the brain itself,
that I call afferent processing.
This processing also creates and/or
strengthens connections back in
the other direction that are crucial
for many higher-level processes.
I call these efferent connections.
|
| 3
|
The recursive application of the above function to incoming data leads to emergent
behaviour that can be described as
abstraction and prediction-enhanced selection.
Only the most important signals make it through this selection process.
|
| 4
|
The end result of many levels of abstraction is the creation/update of symbol schemas,
which are networks of neurons in the brain that represent concepts.
This includes the special self symbol schema that represents “the self”.
|
| 5
|
The total of all the symbol schemas and the connections between them in my brain
is a model of my world, which
includes my body, my brain and its processing.
I use and interact with this model rather than reality.
|
| 6
|
Many high-level brain functions emerge from the existence of symbol schemas and
their connections: reinstatement, attention, perception, cognoception, action and
language. Self-awareness leads to my conclusion that I am my self symbol schema.
|
Brain-wide functions such as
memory and prediction operate
differently at different levels.
Sleep is a very curious brain-wide
feature with multiple purposes.
|
| 7
|
Consciousness is a multi-faceted experience based on this conclusion, and involving
feelings, qualia and emotions (including meaning and pain), attention and memory.
Thought and free will can best be explained at this level.
|
Use scroll bar if necessary to see right-hand side of table →
Level 1 - Physical constituents and their capabilities
- The brain consists of a very large number of special cells called neurons,
a much larger number of one-way connections between pairs of individual neurons called
synapses that facilitate communication of signals between those neurons,
and various support cells known collectively as glia,
all sitting in a jelly-like fluid.
- A neuron has the ability to pass on electrical signals.
- A neuron receives signals from a number of other neurons,
and, if a certain threshold level is reached within a certain time,
it passes on a signal to other further neurons.
- The signal within a neuron that it passes on is an all-or-nothing event,
always a similar voltage, and always in the same direction.
- A signal is passed between a pair of neurons via a synapse.
- Some synapses, when activated, have a positive effect on whether the next neuron
passes on a signal, others a negative one.
- Existing synapses can change in strength and reliability depending on usage,
new synapses can grow, and unused ones may be removed completely.
- The signals within a neuron are electrical, but the majority of synapse connections
in the human brain use chemical signals called neurotransmitters.
- A voltage is maintained across the membrane (skin) of a neuron that allows it to store
a charge and act as a battery.
- When the voltage across the membrane reaches a certain threshold at
a certain place in the neuron cell body, a signal or action potential is created within the neuron.
- This signal is then passed to other neurons via synapses which can contribute to the
electrical charge in those other neurons.
- These processes are all created by the
movement of ions across the membrane.
- The various types of cells known as glia perform important support and
housekeeping roles for neurons and in the brain in general.
- They provide insulation to help the transmission of signals within a neuron.
- They are involved in promoting the growth of the parts of neurons that make
connections, and also assist in the creation and maintenance of those connections.
- They are the equivalent in the brain of immune cells in the blood stream.
- It has only been discovered quite recently that they also have a role in signalling.
- Most of the same neurotransmitter chemicals that are used at
synapses can also be produced
and spread more widely in the brain by neurons and certain types of
glia as well as cells elsewhere in the body (in which case they are normally known as hormones).
This is called neuromodulation, and the effects can be felt by
large numbers of neurons and synapses over a wide area of the brain.
Levels 2 to 5 - Afferent processing of data
- The only external data available to the brain is incoming (afferent) electrical signals from the senses
(both external and internal senses), so processing this data is the only possible way that the brain
can learn about the world, and so learn how to survive.
I propose that data relating to internal brain processes is also analysed in exactly the same way;
this helps the brain manage itself and ultimately leads to self-awareness and consciousness.
- Afferent processing is the application
of common functionality to all incoming data many times over, recursively and hierarchically over
many levels, resulting in increased amounts of compression and abstraction at each level.
- At the lowest level of description,
this common functionality is the combination of coincidence detection, which is a basic
capability of neurons, with the memory capabilities of synapses, and so can be called
memory-enhanced coincidence detection.
Applied over many levels, this provides the compression and abstraction.
- The additional capability of some synapses to create an inhibitory
effect on other neurons needs to be added to this to create the functionality required for
the level 2 processing that I have called
memory-enhanced coincidence detection and lateral inhibition.
- At a higher level of description, when applied recursively and hierarchically,
this functionality creates the level 3 processing that I call
abstraction and prediction-enhanced selection.
- The end result of this processing is many
symbol schemas, described in level 4.
A symbol schema is a network of neurons in the brain that represents a concept, and all
concepts in a brain have such a representation, including those
that represent the processing of the brain itself.
- A special symbol schema representing the self that I call the
self symbol schema
is certainly the largest in the brain, with many components and sub-symbols; it is this
that creates self-awareness and ultimately consciousness (see level 7 below).
- At an even higher level of description, the complete set of symbol schemas
and their connections can be described as
a model of my world. This is described
in level 5.
- The afferent processing of data also creates or strengthens links that
go in the reverse direction, back towards the origin of the data, that I call
efferent connections.
- These are created, or at least strengthened, using the same functions of
memory-enhanced coincidence detection.
- It is only in recent years that neuroscientists have realised how important
these reverse connections are.
- All of the more advanced brain functions in levels
6 and 7 (below)
make extensive use of these efferent connections.
Level 2 - Memory-enhanced coincidence detection and lateral inhibition
- The combination of the signal processing capabilities of neurons
and the variations in synapse connections
outlined in level 1 above
can potentially perform many different and some quite complex operations.
However, three quite simple functions can together be the starting point
to explain all of the higher-level processing described in the levels below.
- Coincidence detection is the most simple and obvious functionality possible
when two neurons each have synapse connections to a third,
which is the basis (the lowest common denominator) of a hierarchical structure.
- In order to easily show how neurons can provide basic coincidence detection,
I have created a very simple model neuron that I call an ABCD neuron.
An ABCD neuron has only two inputs and one output, and generates an output signal when a signal is received
from both inputs at around the same time (this is what I describe as a coincidence).
- This simple functionality is actually part of what a real-life neuron (or set of neurons) provides.
- I have created some afferent processing examples
that show how sense data and internal data from within the brain can be processed using these model neurons.
- If the additional capability of synapses to be able to retain
a simple memory of a coincidence is added, the resulting functionality is
much more powerful.
- A further capability of synapses to create a negative or inhibitory
effect on the ability of a neuron to produce a signal can create what is
sometimes called a lateral inhibition effect.
- The combined functionality of the first two, which can be called memory-enhanced coincidence detection,
not only processes incoming (afferent) data from the senses and from within the brain, but also creates or strengthens
efferent connections back towards the source of the data;
the afferent processing examples also give details on this.
- The third function, lateral inhibition, means that, since many neurons have connections to
other similar nearby neurons, they can mutually inhibit one another. This leads to a form of competition
between parallel processes where only the strongest signals win out. This is the lowest-level basis of the high-level
function of attention (see level 6 below).
- The resulting functionality of all three together can be described as
memory-enhanced coincidence detection and lateral inhibition,
and this is the lowest level of description of afferent processing.
Level 3 - Abstraction and prediction-enhanced selection
- When the process of memory-enhanced coincidence detection and lateral inhibition
(level 2 above) is applied to incoming data recursively over many hierarchical levels,
much more complex and powerful functionality emerges which can be called
abstraction and prediction-enhanced selection. This has three elements:
- Memory-enhanced coincidence detection carried out recursively over many hierarchical
levels leads to emergent behaviour that can be described as compression, extraction of invariance, or abstraction.
The most important, most generic, and/or most common aspects of the data being processed are extracted and stored.
This is the beginnings of creating experience and intelligence.
- Efferent connections, back towards the origin
of the data, created or strengthened across many hierarchical levels are used for prediction,
among other things.
- Lateral inhibition carried out over many hierarchical levels leads to selection of the
most important signals and the suppression of others.
- This selection process can be influenced by the strength of the input signals (bottom-up);
- And/or it can be influenced by the strength of prediction signals (top-down) that use efferent connections.
This overall process of selection is sometimes described as biased competition and is a higher-level description of how the process of attention works
(than the one given in level 2 above).
- The afferent processing examples
give more details on how the first two can happen (the examples do not include the effect of inhibition that creates the selection process).
Level 4 - Symbol schemas and the connections between them
- A symbol schema is my name for a set or network of neurons
and, more particularly, the synapse connections between them which, when activated,
represents a concept in the brain.
- It has been clearly shown in recent years that neurons in the brain do
participate in representing concepts, and these representations have
been given a number of different names.
- Symbol schemas represent concepts of things not only in the outside world, but
also in the body and in the brain itself. There is a symbol schema in my brain for
every concrete and abstract concept of which I have any knowledge, either consciously
or unconsciously, so there must be many thousands of them.
- Any particular neuron is likely to be involved in many different symbol schemas,
so there is a many-to-many relationship between neurons and symbol schemas. It is the
synapse connections that create the network and define the symbol schema.
- Symbol schemas are created or updated as an end result of
afferent processing, the processing
of incoming data (see levels 2 and 3 above).
- They are the end result of the recursive and hierarchical application of
what can be called compression, extraction of invariance, or abstraction.
- Only signals which are considered relevant (salient) enough make it through,
dependant on both top-down and bottom-up influences.
- Top-down influences are predictions, which are influenced by connections
to other near-by or related symbol schemas, and, at this level, are the most dominant.
- Bottom-up influences are mostly non-predicted “surprises” in the
incoming data.
- There is a very special symbol schema
that represents “the self” - the self symbol schema.
- This is created by exactly the same afferent processing
as for all other symbol schemas, but with input data primarily from the body, the internal senses and from
within the brain itself.
- The self symbol schema is a “super-schema”
because it includes many other symbol schemas within itself (there are many other
symbol schemas that are also super-schemas, but they are nowhere near so large).
- The self symbol schema includes symbol schemas that represent brain processes such as
those described in level 6 below and it is this self-referential modelling that creates
self-awareness.
- The activation of a symbol schema involves a large number of the neurons that make up
that schema all firing (generating action potentials) at the same time, or very nearly the same time.
- This means that the signals propagate through the schema rapidly and there is
an element of self-perpetuation: when a critical mass fires, most of them fire.
- There will be mini-circuits within the symbol schema that mean that the
activation can continue for several seconds or more.
- There will also be external looping circuits, particularly when the process
of attention connects a symbol schema to the self symbol schema, which means that there are
external triggers that maintain the activation if it should start to die out.
- On the other hand, when another symbol schema is activated, inhibitory
signals from it may dampen down the external triggers and cause the activation to stop.
Level 5 - A model of my world
- The set of all the symbol schemas and the connections between them in my brain is, in effect, a
model of my world, including my body and my brain, and the relationships
and interactions between all these things. This encompasses all my
memories, experience and intelligence, and is the highest
level of description of
afferent processing.
- Symbol schemas represent all objects, concepts and relationships,
but they are all compressed and abstracted versions of the real thing.
- The many synapse connections between symbol schemas represent connections
and relationships between those concepts in the real world.
- The strength of connections between symbol schemas can be affected by
the end result of neuromodulation.
- My awareness is of this model, not the underlying reality,
so when I perceive something, I am actually perceiving my own model of it,
including its properties, connections and relationships.
- The model includes the body, the position of all parts of the body,
and the position of the body in relation to the rest of the world.
- The model includes the brain and its processes but,
as with all symbol schemas, these are compressed and abstracted versions.
- This also applies to my self, so when I perceive myself, I
are actually perceiving my own model of myself, and yet it is also that
model that is the doing the perceiving.
- The features that emerge at this level are architectural properties
of the model caused primarily by the balance of excitatory and inhibitory synapses.
- This includes its state of criticality, being balanced on the edge of chaos.
- This means that a very small change in perhaps the firing of one neuron
can potentially cause a large change such as the deactivation of one symbol schema and
the activation of a new one.
- This is how my thoughts or attention can flow so easily from one thing
to another, and also means that the workings of the brain are not fully deterministic.
Level 6 - High-level functions that depend on symbol schemas and the connections between them
- The model of my world (including my body and my brain itself), which is made up of
symbol schemas and their connections,
is used by the brain to help it survive in the world. All the high-level functions listed below use this model as well as
the efferent connections from symbol schemas back towards the original sensory neurons or the source of the data.
All these functions are closely interconnected.
- Reinstatement is the term I prefer for the higher level of description of
efferent connections when they relate to the connection of
symbol schemas.
- Efferent connections that are created by afferent processing
are, in effect, a historic record of how a symbol schema was created and updated, because they are paths back
towards the areas that were active at those times. So, at a higher level of description, they keep a trace
of the experiences, feeling and emotions that can be replayed when that symbol schema is reactivated.
- Reinstatement is the way that symbol schemas gain meaning,
through efferent connections back to other symbol schemas.
However, meaning is only manifested as a feeling when the symbol schema is connected to the self symbol schema.
In other words, I am only aware of the meaning of something when I am conscious of it (see level 7 below).
- A useful parallel can be made with the self symbol schema itself: it also gains
“meaning” via reinstatement.
In other words, I have a feeling, or there is meaning, associated with perceiving myself.
- Attention seems to me to be my awareness, when I am awake,
of things that I choose to pay attention to, from moment to moment, in a continuous flow. But actually
what I am aware of is only a schema of the real process of attention, and there is a lot more going
on than I can ever be aware of.
- Attention can be either on external events or on internal memories,
and in both cases is created via reinstatement using
efferent connections from
the self symbol schema to the relevant
symbol schema(s).
These reinstatement routes will be longer for more complex or abstract concepts.
- These looping signals create oscillations,
which is how attention can last for more than a few milliseconds.
These oscillations are often called brain waves, although not all brain waves are created by attention.
- The process of attention can cause incoming data to create or update symbol schemas,
even if the connections do not reach the self symbol schema. In other words, a certain level
of attention can happen without any awareness or consciousness.
- When attention on a particular symbol schema does connect to the self symbol schema,
it is as a result of many levels of lateral inhibition or prediction-enhanced selection
(see levels 2 and 3 above), and
the looping signals enhance this temporary dominance, so only one connection is made at a time,
meaning that only one thing at a time can be conscious.
I can only be aware of, and have attention on, one thing at a time,
although rapid switching is possible.
- Although I feel that I have control over the subject of my attention, in fact
the focus of my attention is selected subconsciously by the process of biased competition
based on signals coming from several directions, only one of which I have any influence over.
Unexpected external events are always likely to take priority.
- My innate feeling is of a continuous flow of my attention from one thing to
another, but in fact the brain is filling in a lot of gaps using
memory and predictions.
- These differences between my innate understanding of attention and
how we now know that it actually works are explained by the proposal that my self symbol schema
can only be aware of a model or schema of how attention works. The symbol schema that represents
attention is built and maintained by the afferent processing of data in my brain about my attention process,
but the only data available to it is from within my self symbol schema, meaning the conscious aspects;
this process is what I call cognoception (see below).
- There are implications in this for free will (see level 7 below)
because attention is the source of all my choices and action.
Attention is not under my conscious control, but I can influence it.
- Perception is not simply the receiving of sense data signals
but is the activation of a symbol schema within the
model of my world, and also therefore
efferent connections back towards sensory neurons
using reinstatement.
- The process of perception involves both afferent sense signals inwards and
efferent prediction signals outwards, meaning that a perception always involves both
understanding and past experience.
- So a perception is the brain’s best guess of what is being sensed.
- The same is true of the perception of internal signals from the body, and
also brain functions that are part of the self symbol schema
(see also cognoception in the following section immediately below).
- Not all perceptions become conscious. Only those signals that make it
through all the levels of possible inhibition within afferent processing
that is part of attention and connect to the
self symbol schema are the ones that become conscious.
- Attention and perception are very closely related:
attention is the signal filtering mechanism,
perception is the activation of a symbol schema by those signals.
- Cognoception is my name for the perception of internal brain data.
- It is an exact equivalent to perception, but instead of the data being from
the senses or inside the body, the data is generated from internal brain processes.
- The symbol schemas created as a result are all part of the self symbol schema,
and play a very important part in self-awareness and consciousness (see level 7 below).
- The processing of data is exactly the same as for afferent processing, but the data available
is only what is available within the self symbol schema, so these symbol schemas only model the
conscious parts of the processes.
- Action is the term used, as in normal speech, to mean the
control of movement and other changes in the body by the brain.
- All action, except basic reflexes, is driven via
efferent connections from
symbol schemas, so action is very closely tied to perception.
- I can only act correctly by using the
model of my world.
- Actions are often driven directly by perceptions, and perceptions are
changed by actions.
- The same method, which could also be called action, is used to control, or
at least influence, brain processes such as attention and decision making.
- The unique ability of humans to use language
is enabled by a special set of
symbol schemas that are symbols for symbols.
- This includes symbols representing written words, which are
symbols for spoken words, which are in turn symbols for other concepts.
- Language gives us a big advantage in being able to more easily
define abstract and hierarchical concepts, by creating symbol schemas for this concepts.
- Many other fields that involve abstract concepts also make
use of meta-symbols, such as mathematics and music.
- Self-awareness is the underlying requirement for
consciousness in level 7 below. It means that the brain is capable of thinking
about itself and its own processes.
- Self-awareness is generated in the
self symbol schema primarily as a result of the modelling of the
process of attention.
- The processing of data from within the brain (that I call
cognoception) relating to the process of attention, using
exactly the same hierarchical afferent processing that is used for
incoming sense data, creates a symbol schema that represents the process of attention.
- When this is activated, it represents not only my attention on a particular other
symbol schema, and therefore my perception of the object or concept that this represents,
but also my awareness of my attention on that other symbol schema.
- If self-awareness is my self symbol schema being aware of itself,
it means that what I think of as “I” is the same as my self symbol schema. My conclusion that
I am my self symbol schema has many implications, including making
explanations of consciousness and feelings much easier.
Brain-wide functions that operate at different levels
- Memory and prediction are very closely connected, but
sleep would seem to be unconnected to anything else, although it is a brain-wide phenomenon.
- Memory is brain-wide function that means
different things at different levels, but in all cases it involves physical changes either to
synapses or within neurons.
- At the lowest level, it could be an accumulation of some protein molecules
or charged particles (ions) in a particular place or places in one or more neurons
that only lasts for a few milliseconds at most, but in that time could contribute
to a neuron firing, or not firing. This could be a small part of a decision-making
process involving prediction (see immediately below) and could help decide
what to pay attention to or what action to take.
- At the highest level, it could be permanent connections consisting
of many synapses between many neurons in different symbol schemas that can last a lifetime.
This could represent a long-term cherished memory, an important practised movement,
or be a crucial part of my personality or temperament.
- Between these two extremes are the memories that I am most aware of
on a day-to-day basis that involve connections within and between symbol schemas that
last for seconds, minutes, hours or days. When activated by the process of
attention they enable me to remember, for example, what I did last week
(episodic memories), and what I am due to do today.
- My own knowledge of my own memory process is only via a memory schema that is
built in my brain by cognoception, which is the afferent processing of internal data about
my memory processes.
- Prediction is a brain-wide function that has a
part in many levels of description.
- Any memory, at any level (see immediately above), can contribute to a prediction.
- All prediction is driven by a memory, but not all memory is used as a prediction.
- Prediction can help to process a large amount of incoming data more quickly.
- It can also assist with the processing of incomplete data by filling-in the gaps.
- Perception uses prediction just as much as it uses incoming sense data,
so prediction is a fundamental part of how I perceive the world.
- Prediction plays a crucial part in most every-day activities such
as walking, running, cycling, picking up an object, playing sport, and so on.
- Some recent proposals describe the brain as primarily a prediction machine,
and say that this view can explain not only perception,
but also action
and attention.
- Sleep is a brain-wide function that is a requirement for
the brain to work properly. It has a number of important functions, not all of which are fully understood yet.
Some important ones are the maintenance of the balance of excitatory and inhibitory connections between neurons
and therefore also between symbol schemas, the clearance of waste products from the brain, and the consolidation of memories.
Level 7 - Consciousness, feelings, thought and free will
- My experience of consciousness is multi-faceted,
based on self-awareness and
my conclusion that I am my self symbol schema, but also comprising
attention, memory and
feelings, qualia and emotions.
- What I am aware of, from moment to moment, is the result of the process of
attention (see level 6 above), but modulated by my model of my
process of attention.
- Memory (see brain-wide functions above)
is a requirement for full consciousness.
- If I have consciousness but no memory, I may feel conscious at any
particular moment, but my consciousness has very little useful purpose.
- One of the main reasons that consciousness evolved is to
provide an episodic memory, because it creates a considerable survival advantage.
- Feelings, qualia and emotions are a unique feature of
consciousness that are not experienced without self-awareness, and I include “meaning” and pain in this as well.
- Each specific feeling or emotion is represented by a symbol schema
and can be connected to another symbol schema to represent that
symbol schema being associated with the feeling or emotion.
- The qualia of an object or event are feelings or emotions linked to a symbol schema
that are activated when that symbol schema is activated.
- The subjective experiences (phenomenality) of feelings, qualia or emotions
are generated by the reinstatement of sensory neurons
that were activated when the symbol schema was created or updated.
- Feelings are manifested in the self symbol schema
when a connection is made, via the process of attention, from it to the
symbol schema in question.
- When this happens, I consciously perceive that feeling or emotion,
in exactly the same way as I consciously perceive any other concept or object.
- Recalled feelings, qualia and emotions are a special form of
reinstatement using
efferent connections from
the self symbol schema back towards the “sensory”
areas that created those symbols, which for emotions can be internal parts of the body.
- For emotions, this communication may not be solely via neurons or
synapses but via the class of
neuromodulators known as hormones.
- Thought, along with imagination, planning, day-dreaming and
also night-time dreaming are all aspects of the same thing.
- They involve the activation of symbol schemas
that may (or may not) become conscious at least to some extent.
- When these activations are conscious, they become so through the process of attention.
- This also includes when movements are “rehearsed” but not actually
carried out, for example when
watching someone else doing something, or when reading “internally”.
- Free will is also modelled
within the self symbol schema and is a symbol schema in its own right.
- It comes about from modelling the ability to activate symbol schemas that
represent attention, which means my choice of what to think, say or do.
- The choices I make are influenced by my past experiences, my preferences,
my beliefs, desires and goals, all of which are represented by symbol schemas and their connections.
- My self symbol schema can influence the process of attention by the
proven process of downward causation.
- This means that I do have the ability to at least influence my choices of
what I think, say or do, but it is not as clear cut as my model of free will leads me to believe.
Conclusions
- The ideas presented here are self-referential:
- What I have described here is essentially a model of how the human brain works.
- Arguably the most important element of how the brain works is that
it builds a model of my world, including itself, by combining representative symbols.
- In working on this, I have been using my own brain to read books, articles and research
and to think about and try to understand how the brain works.
- I therefore have created new symbols and a model in my brain of how the brain builds
and uses a model.
- If you have read and understand at least some of what I have written, your brain
now contains symbols of some of my ideas and a model of my model of how the brain builds and uses a model
(see mind-boggling brain!)
Page last uploaded
Mon Feb 19 08:20:51 2024 MST
