Scientific Understanding of Consciousness
Introduction to Consciousness
Consciousness is an emergent property of a biological network of neurons resulting from billions of years of evolution.
Human brains have evolved over hundreds of millions of years from pre-human animals that have been successful in their struggle for survival. Each of us has a brain that is a result of the genetics of our ancestors together with the unique neurological development, both prenatally and postnatally, our brains have experienced in the environment. The intricate mechanisms in our synapses are the result of billions of years of molecular and cellular evolution. Our consciousness and our uniqueness as an individual person are stored in the functionality of more than 100 trillion synapses in our brain.
Consciousness not precisely defined
The term consciousness is not precisely defined, nor should it be at the current status of the science. We should not nitpick words and definitions in current descriptions relating to consciousness. If we get the gist of an idea being alluded to, that should be sufficient at the present time.
Clarification of the nomenclature is typically the first sign of maturity for scientific enterprise. (Andersen; Hippocampus Book, 42)
Human-type Consciousness founded upon Core Consciousness
Each of us has human-type consciousness, a huge expansion of the core consciousness usually associated with mammals such as dogs. I discuss core consciousness and human-type consciousness in separate sections.
Consciousness: ‘convolution’ of a current mental image with sense of self
Consciousness is a ‘convolution’ (my choice of word; it's hard to precisely convey the concept) of a current mental image with a neural network image representing the self. Edelman uses the term ‘remembered present’ to express this idea. ‘Remembered’ conveys the notion of an accumulated store of individualized synaptic functionality representing the self. ‘Present’ conveys the notion of a current mental image, which can result either from sensory input data, or from a [reconstruction] of some past memory, or from some imagined thought. Working memory in the prefrontal cortex is likely to invoke neuronal network activity in other brain areas to synchronize into cyclical thalamocortical activity comprising the dynamic core. The images generated by fMRI studies give some indication of the brain activity that may include the dynamic core involved in various tasks.
Sense of Self
A person's individuality and personality are stored in the functionality pattern of many billions of synapses whose connectivity is determined by genetics and the environment in which the brain develops. The total synaptic ensemble comprising the self is established over a lifetime of experience, with the detailed connectivity changing dynamically as a person experiences the environment. Only a small subset of the brain’s 1015 possible synapses actively represents the person's sense of self at a given moment. (e.g. sometimes a pain may be in the forefront; fortunately, most of the time it is not.) Keep in mind that the no specific set of neurons is required for the sense of self. Only the pattern of synaptic connectivity is preserved; individual neurons representing this connectivity pattern can vary.
Consciousness requires the neuronal representation of the self (Antonio Damasio).
Thalamocortical activity, supported by subcortical activity, mediates a neural representation of a worldly object or concept, either perceived or imagined. This neural representation is called a “mental image.” This momentary ‘mental image’ representing a thought is ‘convolved with’ the neural representation of ‘the self’ to yield core consciousness.
Gestalts in Neuronal Assemblies
Mental images are likely to be formed from gestalts of neuronal assemblies in the neural network. I enjoy experiencing my brain forming gestalts as I watch cloud formations in the sky when fracto-cumulus clouds break up into cloud-free fair weather. I have fun visualizing lambs, puppies, old men with beards, etc.
Complex mental images are likely to be formed from hierarchies of gestalts. I believe the brain forms gestalts comprised of assemblies of neurons recursively linked via synapses.
Our waketime daily lives are a continuous sequence of perceptions. In my understanding, perceptions are comprised of hierarchies of gestalts. Non-conscious background neural activity includes thousands of gestalts comprised of independently circulating neural signals. As perceptions are formed, hierarchies of gestalts are recruited by the associative property of memory into ever larger nested hierarchical neural assemblies with reentrant recursion among functional areas of cortex and subcortical areas. This interactive neuronal network activity produces a wide-spectrum EEG signal with prominent activity at ~40-Hz. Some portion of this network activity is likely associated with the ‘dynamic core’ of consciousness. As we all know, perceptions (thoughts) change on the basis of about half a second or less.
Bayesian Inference and Recursion
Without getting too technical, let me just alert you that much of the brain's neuronal activity can be interpreted as an implementation of the statistical technique of Bayesian inference with recursion. If you are inclined to have an interest in these topics, you may be interested when I mention them in further discussions.
Consciousness mediated by the Dynamic Core
My current hypothesis is that consciousness is mediated by a dynamic core (Gerald Edelman) of thalamocortical activity, which is ever changing on the basis of about 10 ms. Neurons comprising the dynamic core are ever changing. A particular thought at one instant can be comprised of a connected network of some hundreds of millions of neurons. If thoughts change and moments later return to the same thought, the individual neurons representing the thought will likely be slightly different. Thus the network comprising a particular thought need not require a precisely determined network of neurons. Individual neurons will sometimes (often fleetingly) participate in the dynamic core, but most often will not be a part of the dynamic core. Individual neurons comprising the dynamic core can be reused and connected differently 10 ms later as the dynamic core changes to mediate a different thought. The size of the dynamic core and the number of neurons involved can change dramatically as the intensity of thought varies from reverie to intense cogitation. Each instantaneous dynamic core of consciousness is likely to involve some neurons in the prefrontal cortex constituting working memory.
Patterns of neural signal traces in the dendritic trees of neurons, dynamically connected momentarily by efficacious synapses, sculptured by genetics and experience, mediate neural network activity, an ever-changing subset of which forms the dynamic core of consciousness.
Minimal dynamic core required for consciousness
The minimal dynamic core required for consciousness is currently unknown, but may become more clearly understood in the future decades. Uncertain consciousness states such as Persistent Vegetative State (PVS) continually present a challenge for the clinical medical community.
Modularity of brain anatomy and functions
The brain is modular in a number of ways. Evolution has produced a distinct anatomical modularity in the brainstem, subcortical structures, and cortex. All of these structures have intricate anatomical connectivity and functionality. Here are some of the ways the brain's modularity can be considered:
Evolutionary and anatomical modularity of brain -- Brain Stem, Limbic System, Cortex -- approximately corresponding to reptiles, mammals, and humans in the evolutionary sequence.
Four lobes of cortex -- frontal, parietal, occipital, temporal.
Approximately 50 Brodmann areas of cortex.
Functional areas of cortex -- Boca, Wernicke, visual hierarchy, auditory, somatosensory, motor, hippocampus, prefrontal, cingulate, etc.
Neocortex has six fairly distinct levels.
Paleocortex such as hippocampus has three levels.
Cortex has modular columns extending through the six levels.
Fortunately, the brain's overall functionality can be considered in terms of three topological networks working closely together
Three Topological Networks of the Brain
The brain's neural network can be considered in terms of three topological networks, closely interworking together. The thalamocortical system forms the fundamental biological mechanism of consciousness, supported by other functionality. A set of parallel, unidirectional neural pathways link the cortex to a set of its appendages, each with a special structure -- the cerebellum, the basal ganglia, and the hippocampus. Fan-out meshworks of diffusely projecting neuromodulatory neurons emanate from brain stem and midbrain nuclei. These small collections of neurons can deliver a dose of dopamine, norepinephrine, serotonin or acetylcholine to widespread regions of the brain including the cerebral cortex and basal ganglia.
The thalamocortical system consists of continuously-ongoing, reentrant neural signals between nuclei in the thalamus and neurons in all modular areas of the cortex. In general, neurons are never idle. Neurons fire at about 5 Hz in the quiescent state and at rates of 100 Hz or more in the active state. A subset of thalamocortical activity comprises the dynamic core of consciousness. The magnitude of the dynamic core varies greatly with the intensity of thought.
The thalamocortical system mediating the dynamic core becomes active early in embryonic life and continues to be active until death or until trauma states such as brain death.
Parallel System through Basal Ganglia, Cerebellum, and Thalamus
A set of parallel, unidirectional chains link the cortex to a set of its appendages, each with a special structure -- the cerebellum, the basal ganglia, and the hippocampus. Cerebellum is concerned with the coordination and synchrony of motion, although it is also involved in aspects of thought and language. Basal ganglia are involved in the planning and execution of complex motor and cognitive acts. Hippocampuses, which are evolutionarily ancient structures deep inside the temporal lobes, are involved in the process whereby short-term memory gets transferred to long-term memory, and also spatial memory.
Diffusely Projecting Modulatory Neurons
Fan-out meshworks of diffusely projecting neuromodulatory neurons emanate from brain stem and midbrain nuclei. Reentrant circuits of the thalamocortical system are modulated by these neurotransmitters. A number of amine chemicals are associated with states of arousal. Four of them are: (1) serotonin, (2) acetylcholine, (3) dopamine, (4) norepinephrine.
Human brains have billions of neurons that together make trillions of synaptic connections among one another. The overall stochastic nature of neuronal behavior suggests that the physiologically meaningful signal from cortex should be the average firing rates of a. population of perhaps 100 to 1000 neurons near a particular cortical site.
Neurons and Synapses
Neurons communicate information via neurotransmitters in synapses. A neuron’s thousands of dendritic synapses receive pulse-like input from other neurons. Whenever a large group of spikes arrives nearly simultaneously, a neuron will generate an action potential, which is transmitted down its axon. This near simultaneous arrival of input spikes creates pulse-like behavior.
Neurotransmitter molecules diffusing across synaptic space between neurons mediate the transmission of information in the neural network. Neurotransmitters mediating the rapid flow of information are released from pre-synaptic terminals of one neuron and diffused in a millisecond or so to the post-synaptic receptor of another neuron. Modulatory neurotransmitters, operating in seconds, minutes or longer, often emanate from widely-projecting neurons clustered in sub-cortical ganglia.
Modulatory neural activity. Neurotransmitters.
Dendritic Trees and Stochastic Neuronal Behavior
Projection neurons typically have a dendritic tree with perhaps 10,000 synapses. Because neurons signals occur in the form of pulses of about 1 or 2 ms, a population of synaptic input signals on a neuron' s dendritic tree must combine in any given time interval of the few milliseconds to exceed the neuron’s threshold to cause the neuron to fire. The result is a stochastic firing behavior for an individual neuron. Groups of closely allied neurons tend to fire together resulting in a more deterministic behavior. The population-based functionality of neuronal groups means that specific individual neurons are not required, and the functionality of the neural network can survive the death of the few neurons. It also means that when synaptic patterns of memories are reactivated, a slightly different pattern of synapses will be involved for each memory activation.
The wide range in the power spectrum observed in the EEG can be considered the result of reentry and recursion among nested hierarchical levels in the neural network. The synchronizations observed are the result of coherence in the neural computations rather than a cause of the network functionality.
Simultaneity of neuronal activity arising from intrinsic oscillatory electrical activity, resonance, and coherence are at the root of cognition. (Llinás; I of the Vortex, 12)
Rhythmic activity can be generated by two main mechanisms -- intrinsic membrane properties and synaptic circuits. (Shepherd and Koch; Synaptic Circuits, 30)
Oscillations have been documented in the brains of numerous mammalian species, ranging from very slow oscillations with periods of minutes to very fast oscillations with frequencies reaching 600 Hz. (Buzsáki - Rhythms of the Brain, 112)
The neural network likely functions as a nested hierarchy of recursive loops operating at perhaps ~20 ms for visual sensory circuits, maybe ~50 ms in association cortex areas for perception, and perhaps >1 second for decision circuits in frontal cortex.
Brain oscillators should be considered as a system of oscillators with an intricate relationship between the various rhythmic components. (Buzsáki; Rhythms of the Brain, 119)
In general, slow oscillators can involve many neurons in large brain areas, whereas the short time windows of fast oscillators facilitate local integration, largely because of the limitations of the axon conduction delays. (Buzsáki; Rhythms of the Brain, 115)
The amplitude of the EEG power spectrum increases as the frequency decreases. This inverse relationship is expressed as the "one over f" power spectrum (also called "pink" noise). (Buzsáki; Rhythms of the Brain, 119)
Plasticity of Neural Connections
The configuration of the neural network is continuously in a state of flux via the plasticity of synaptic connections. The LTP (long-term potentiation) of Hebbian plasticity, which hypothesizes that ‘neurons that fire together wire together’, begins the network formation prenatally and continues with lifelong changes that modify the sensitivities of the brain’s 1015 synapses.
Plasticity of Synapses Mediates Memory. Biochemical changes in the synapses mediate memory in the relative near-term, whereas gene expression and protein changes in synapses and dendritic tree structures are consolidated over time for hippocampus-independent long-term memory. Neural signals activate a widespread but sparse memory trace that most closely conforms to a synaptic efficacy pattern established by prior neural activity.
Memory -- Declarative, Procedural, Emotional, and Working Memory
Memory can be parsed into several types, including declarative memory, procedural memory, emotional memory, and working memory. Edelman's idea of consciousness as ‘remembered present’ implies that distant-past (not via hippocampus) recall of declarative memory must be functional, as exemplified in the case of epilepsy patient H.M. Procedural memory is [probably] not required for consciousness.
In the real world, multiple streams of information reach our awareness, some of it relevant, some not for the task at hand. With the inherent capacity limitations of working memory, it is essential that only representations of task-relevant information are generated and maintained.
Human interaction with our environment involves a fluid integration of externally driven perceptual information that demands attention based on stimulus salience or novelty (bottom-up processes) and internally driven, goal-directed decisions concerning external stimuli or stored representations (top-down modulation).
Fear – Pleasure
Release of dopamine onto the nucleus accumbens appears to underlie all reward feelings. The dopaminergic projection from the ventral tegmental area (VTA) to the nucleus accumbens is a key feature of the pleasure circuitry.
Emotions are among the very oldest of the brain's properties. Limbic system, hypothalamus, and brain stem intervene in body regulation and in all neural processes on which mind phenomena are based. Emotions are linked to the motor aspects of FAPs by access through the amygdala and the hypothalamus and their connectivity with the brain stem.
Although there is no consensus among experts, six so-called primary or universal emotions are often stated, for example: Happiness, Sadness, Fear, Anger, Surprise, Disgust.
The limbic system consists of a number of subcortical structures which are active in emotions and which are tightly interrelated with cortical functions. In recent years, some neuroscientists have included the prefrontal cortex as a member of the limbic system, since it is often heavily involved in emotional activity.
As I view it, motivation to action arises from a range of neural signals ranging from (1) fear of a perceived threat to (2) groping for an expected pleasure. Joseph Ledoux states that interactions between the amygdala and nucleus accumbens contribute to motivation.
Motivational circuitry of the brain includes the prefrontal cortex, basolateral amygdala, and ventral pallidum. The dopaminergic projection from the ventral tegmental area (VTA) to the nucleus accumbens is a key feature of the circuitry.
The motivational circuitry of the brain connects the prefrontal cortex decision process to the premotor cortex, which uses its FAPs along with FAPs stored in the basal ganglia and cerebellum to plan movement, followed by the motor cortex, which invokes FAPs in the brain stem and spinal cord to produce movement.
Movement control is nearly always the functional result of wake-time brain activity. In the course of normal conversation, for example, the brain activity of listening to a friend and composing your own thoughts is then followed by the muscular movements of vocalization of speech, facial gestures, and gesticulations of arm and body movements.
Movement is facilitated in the brain and nervous system by a hierarchy of modular functionality I call FAPs but known by a number of names, including gestures, synergy, schemas, motor programs, stereotypical patterns of movement, etc.
FAPs are most probably implemented at the level of the basal ganglia and put into context by the reentry of the basal ganglia output into the ever-cycling thalamocortical system.
In addition to voluntary movement, a person can experience involuntary movement in response to a startle reaction. For example, a visual startle pathway goes via the subcortical structure of the superior colliculus to invoke motor FAPs of muscular body movement.
Friendly conversation is a brain activity that involves much of the brain's functionality of everyday consciousness, although language is probably not required for minimal consciousness. Visual and auditory senses provide input data for the language functionality. Cognition and emotion provide links to autobiographical memory and, via a neuronal process of Bayesian statistical inference, a continuous stream of new perceptions. Premotor cortex, basal ganglia, and cerebellum provide motor planning for all of the non-conscious movements and FAPs of voice, gestures, and emotional reactions. Motor cortex invokes FAPs in the brain stem and spinal cord to produce movement for vocal speech along with gestures and facial expressions.
Example scenario of normal consciousness — Unexpectedly Meeting a Friend
Language and Consciousness
Language is probably not required for consciousness. (Zeman; Consciousness, 285)
We have mirror neurons scattered throughout key parts of our brain that fire as we perform an action and also fire when we watch someone else perform the same action. (Horstman; Day in Life of Your Brain, 51)
Mirror neurons are found in areas associated with movement and perception, as well as in the regions that correspond to language and understanding someone else's feelings and intentions: the pre-motor cortex, the inferior and posterior parietal lobe, the superior temporal sulcus, and the insula. (Horstman; Day in Life of Your Brain, 52)
Low mirror neuron activity is common in people with autism, which is thought to be due in part to flaws in the mirror neuron system. (Horstman; Day in Life of Your Brain, 52)
Sleep and dreaming -- Memory Consolidation – Creativity
Experts hypothesize that memories are consolidated during sleep and dreaming. Dreaming is an altered state of consciousness, akin to those induced by psychedelic drugs in waking.
Complexity, Self-Organization, Emergence
Complex systems involve nonlinear interactions between a large number of simple elements. Self-organization reflects matter's incessant attempts to organize itself into ever more complex structures, even in the face of the incessant forces of dissolution of the second law of thermodynamics. Molecules tend to coagulate in such a way that their atomic faces bind or "fall together," forming a lower energy state conforming to the second law of thermodynamics. This is a microscopic example of the macroscopic observation that "rocks fall down hill."
Emergence reflects the incessant urge of complex systems to organize themselves into patterns.
Metaphors can sometimes help
A metaphor or a simple example, such as rocks falling downhill as an example of the second law of thermodynamics, can sometimes clarify arcane concepts. I'll occasionally use metaphors where I think they may be helpful and not introduce additional confusion.
Further discussion -- Covington Theory of Consciousness