Physiology F6 - Sleep - Principles of EEG Recording PDF

PHYSIO F6 – Principles of EEG recording and sleep 1.​ Principles of the EEG The EEG or Electroencephalography is a non-invasive technique mainly used to record the scalp activity and more practically, for instance, during a running activity, to record the different parameters, or in epileptic p...

PHYSIO F6 – Principles of EEG recording and sleep 1.​ Principles of the EEG The EEG or Electroencephalography is a non-invasive technique mainly used to record the scalp activity and more practically, for instance, during a running activity, to record the different parameters, or in epileptic patients as well. Hans Berger was a psychiatrist at the University of Jena and the first one recording scalp activities in humans (1929), after intracellular recordings, or still more invasive techniques, were already performed several years before with the EEG being considered as clinically non-relevant. In order to be able to elaborate this electric signal, more than one device is needed. A special cap is applied on the head of the patient to record cortical brain waves. The so recorded cortical brain waves are sent to an amplifier/ADC, which then makes it available for a computer to finally obtain the conversion into a digital signal. ​ The EEG is basically the graphical representation of a potential difference between a recording “active” electrode (REC), placed on the skull surface (ideally represented by the orange stripe below, composed by many firing neurons), and a reference indifferent electrode (REF), placed on another body surface. It is a dynamic measure as the potential difference is represented as a function of time. The main physiological mechanism generating the actual recorded signal is the activity of cortical pyramidal neurons, highly performing in this case not only as big units, but also for their consistent dendritic prolongments able to run till the first layer of the cortex. They are oriented perpendicularly to the cortical surface. The EEG is thought to be primarily generates by the cortical pyramidal neurons in the cerebral cortex that are oriented perpendicularly to the brain’s surface. The neural activity is detectable by the EEG is the summation of the excitatory and inhibitory postsynaptic potentials of the relatively large groups of neurons firing synchronously. ​ In the picture below on the left, a pyramidal neuron with its soma and its dendrites pointing towards the cortical surface can be observed. When the excitatory post-synaptic potential (EPSP) on the apical dendrite of this cortical pyramidal neuron receives an input by a contralateral cortical afferent, a positive charge enters the dendrite at the site of the EPSP generation, creating what is commonly called a ‘current sink’. This charge forms a loop, known as ‘dipole’, that goes down through the soma of the pyramidal neuron and gets out of it, still completing the loop, at a specific spot called ‘current source’. The EEG basically records many of these phenomena occurring simultaneously and under the electrodes’ measurements. The current measured at the level of the current sink and the one measured at the current source have opposite polarities since positive and negative charges are displayed in a different fashion: negative charges are recorded around the current sink, whereas the positive ones do have a greater affinity for the current source region. The EEG signal is based on the flow of ionic currents generated by the neurons in the extracellular space and this, of course, means that it is not occurring inside the cell. However, the previous evidence can undergo some changes according to the type of afferent signals reaching the pyramidal neuron. For instance, signals coming from the thalamus hit the pyramidal neuron inside layer IV, closer to the soma and so around the current source, whereas signals coming from the corpus callosum hit the neuron more superficially in the cortex. In the case of an EPSP generated by thalamo-cortical afferent, the inversion of the electrical potential occurs and the direction of the current points toward the positively charged dendrites. The polarity of the signal depends on where the synaptic activity is localized within the cortical layers. Thus, this is the representation of a pyramidal dipole, the neurophysiological unit able to generate an EEG signal. There are many dipoles undergoing this mechanism at the same time and so, what the electrodes end up with is a final wave, resulting from the summation of all the recorded changes in polarization. 2.​ Factors affecting the amplitude of the signal The distance between the recording electrode and the reference one plays an important role in influencing the amplitude of the signal: the more distant they are from each other, the bigger the amplitude of the signal grows. This occurs in a proportional way, along with the increasing number of neurons below the electrodes. Another factor influencing the amplitude of the signal is the intrinsic characterization of the cortex. It is not smooth because of the numerous sulci present on its surface leading to the dipoles being not always perpendicular to the scalp. In the case of a neuron perpendicularly reaching the cortex, the amplitude is bigger, whereas in the case of a neuron heading either to the crown of a sulcus or to its bottom, the amplitude is then much smaller. In addition, in the case of two electrodes recording from the opposite sides of the bottom of the sulcus, the result is a neutralization of the signal. Still, one more factor affecting the amplitude of the signal is a physiological phenomenon known as synchronization. In the case of the six neurons of the example below receiving afferents allowing them to fire in a synchronized way, the summation of the six activities results in a final wave increased in amplitude. In the opposite case, the resulting wave is very irregular and small in amplitude. A metaphor as the one of the stadium could be helpful to fully understand the phenomenon. When people (in the metaphor standing for the human firing neurons) are exulting and screaming in a non-synchronous way, a clear sound cannot be distinguished, and everything results in confusion. Instead, when people exult together and in a synchronous way, creating a chorus maybe, the final sound is pretty clear, stronger and can finally be understood. 3.​ EEG principles Although the action potentials, as the larger electrical potential generated by the neurons, may appear to be the most obvious source of the electrical potentials recorded from the scalp, they contribute minimally to the genesis of EEG graphoelements for two reasons. The first one is that the amplitude of the electric field produced by the propagation of an action potential decreases much more rapidly than the amplitude of the fields produced by the postsynaptic potentials in the extracellular environment. The second reason is that the duration of action potentials is very short (1ms), and insufficient to obtain an adequate synchronization of large cortical neuronal populations. So, the action potentials minimally contribute to the EEG signal which is, instead, mainly generated by the extracellular activity and, at specific frequencies, also by either inhibiting or excitatory postsynaptic potentials, since these affect the spiking activity of the neurons. The EEG signal also depends on whether the electrodes are applied externally and so separated from the cortical surface by the meninges, or directly on the cortical surface (intracortical electrodes). Conversely, the flows of synaptic currents in the extracellular space last for about 10–40 ms, so the postsynaptic potentials can add up together more efficiently than the action potentials, and create electric fields large enough to be able to be registered from outside, even without a perfect synchronization. Moreover, the excitatory or inhibitory postsynaptic potential show other peculiarities that make it different from the action potential and more suitable for the EEG signal and graphoelements to be recorded. It does not have a specific threshold and is local, this means it has no tendency to propagate without decrement (as the action potential), but it decreases as moving away from its source. These are so two different features for the post synaptic signals to be more suitable for the EEG signal recoding than the action potential. In the picture below, other techniques are shown. -​ MEG is quite similar to EEG but is able to record and elaborate the magnetic field generated by the electrical activity of the neurons instead. -​ Single or Multi-unit activity (SUA or MUA) are commonly used in animals and imply the electrode measuring very close to the soma of the pyramidal neuron, obtaining a discrete event still leading to the action potential. -​ LFP, or local-field potential, is an intracortical technique, like the previous one, but can record the activity of a huge number of neurons in the extracellular environment. In fact, the graphoelements of the LFP technique appear more similar to the ones of the EEG, but with a higher cortical resolution. -​ ECoG is a sort of EEG but instead of applying the electrodes on the skull, they are applied directly on the cortical surface. This way a wider spectrum of frequencies and amplitudes can be recorded with regards to the EEG. As it can be appreciated in the picture, proceeding along with the different techniques from left to right, they decrease in invasiveness, decrease in spatial resolution, but increase in the size of the neuronal cluster recorded. 4.​ Different ways of recording the EEG activity The first technique to record the EEG is the bipolar recording, when the recording electrode and the reference one are very close and the voltage difference between them is detectable. In this case there is a very high spatial resolution when compared to the monopolar recording, in which the reference electrode is distantly positioned (the ear lobe for instance), but still useful when detecting properties of different areas. ​ In the picture below, the difference between a direct electrophysiological technique, such as the EEG, and an indirect one, such as the functional MRI (fMRI), can be observed. The fMRI is based on the bold signal which is related to the oxygen level-dependent activity in vessels close to the neurons, therefore it is an indirect way of recording the cortical activity. The great advantage of the EEG and the other direct techniques is the temporal resolution, this means that anything happening among the neurons is immediately detectable. The bold signal, being an indirect recording, takes more time and is given by higher and prolonged stimulations. In fact, in any case the bold signal is delayed with regards to the stimulus, but in the specific case this happens to be too small, it could even be undetectable and result in no bold signal at all. The evoked potential, instead, is much different, since the delay is still present but at a physiological level and consists of the specific stimulus taking its time to reach its target. But still, the main advantage is that even a small stimulus can induce a response in the evoked potential. Different waves contribute to the normal brain activity: the one recorded in a normal awake state, with a wave range of 1-20 Hz in one second. This apparent restricted amplitude is given by the mixing up of different signals, each of them having a specific physiological meaning. In fact, performing a Power Spectrum, which is a frequency-domain analysis, different peaks of different amplitudes at different frequencies are detectable: -​ Δ, -​ theta Θ, -​ α, -​ β, -​ γ. The amplitude of these peaks decreases along with an increasing frequency.  Normal activity: apparent low amplitude due to the mixing of all signals present during wake-fulness. Separate signal groups can be separated via spectral analysis. The first wave classes contributing to the final normal activity graph are the ‘low activity’ ones: the delta Δ waves (range of 0-4 Hz) and the theta Θ waves (range of 4-7 Hz). They characterize a specific brain state, the sleeping one. The theta waves are associated with drowsiness or arousal in older children and adults and with the inhibition of elicited responses where a subject is actively trying to repress a response or action and found in young children in all wakeful states. The delta waves are even slower and higher in amplitude, normally seen in adults during slow wave sleep and in babies that are both awake and asleep. Sometimes this has also been found during continuous attention tasks. ​ Increasing the frequency ranges (8-12 Hz), there they come the alpha α waves and the Mu µ rhythms. The alpha waves, also known as Berger’s waves, are mainly recorded at the level of the occipital cortex and arise from the synchronous and coherent (in phase and constructive) electrical activity of thalamic pacemaker cells. They are associated with relaxed and reflecting states of mind, closing the eyes with the control inhibition. In fact, when closing the eyes, there is a huge synchronization of the alpha activity. The Mu µ rhythms are recorded in the perirolandic region, close to the primary sensorimotor cortex. They are believed to reflect the electrical output of the synchronization of large portions of pyramidal neurons of the motor cortex which control hand and arm movement when inactive. Desynchronization can occur during movements by subjects as well as when viewing those movements in someone else. Summing up, the Mu rhythms are synchronized in the resting phases and desynchronized during movements. They are a marker for movements since these waves lose their synchro according to them. Increasing the frequencies and decreasing the amplitudes once more, there are the beta β waves (12-30 Hz) and gamma γ waves (25-100). Beta waves are usually split into three bands by frequency: high (18-30 Hz), beta (15-18 Hz), and low (12-15 Hz). Normally, they are associated with normal waking consciousness, busy or anxious thinking and concentration, and during the inhibition of impulsive behaviors as well. The gamma γ waves are the highest in frequency usually starting from 30 Hz, therefore it is a band verry difficult to analyze because of the very small amplitudes and so the ECoG, for instance, could be more efficient in this case, since it records directly from the cortical surface. About the gamma waves, they are very interesting within the whole EEG spectrum and a popular theory claims that these rhythms represent the binding together of different populations of neurons into a network with the purpose of carrying out a certain cognitive or motor function. They are normally associated with short term memory matching of recognized objects and during sensory processing that involves different senses such as sight and sound. Additionally, several studies prove that an increase in the gamma band may be directly associated to the spiking activity. An American scientist, after performing ECoG and SUA simultaneously, found that the spiking of neurons was associated to an increase of the gamma band amplitude in synchronized way. 5.​ Other waves The human EEG recording is characterized by many other waves as well. For instance, it is useful to study the ‘spikes’, being a typical epileptic signal with a sharp fashion. [He didn’t go into detail saying we are having a practical activity specifically on this]  6.​ The different applications of the EEG signal The EEG signal is also used to detect an evoked potential, and so the differences in potentials caused by a specific discrete stimulus. The visual evoked potential can be taken as an example. There is a specific clinical test, performed to verify the integrity of the visual pathways at its different steps, and so from the eyes to the thalamus, and from the thalamus to the cortex. The main activity is finally recorded in the region of the occipital lobe. Link to the clinical test https://www.youtube.com/watch?v=iXXxL0EOJqs From a clinical perspective, it is important to be able to detect and understand latencies and unusual amplitudes of a peak that could result from a damage along the pathway, affecting the firing rate of the neurons at a cortical level. Usually, the latency stands for the time that the visual stimulus takes to reach the cortical target. SLEEP EEG has been very useful to instigate the stages of sleeps as well. I n 1953 Nathaniel Kleitman and Eugene Aserinksy showed, by means of electroencephalographic recordings from normal subjects, that sleep actually comprises different stages that occur in a characteristic sequence. The EEG technique has been fundamental for an experimental setup to understand the biological meaning of sleeping. In this famous experiment specifically, the EEG has been used to trigger a movement of the cage of the brown rat. This way the rat obviously woke up as soon as it just fell asleep and was not allowed for a proper rest. Consequently, several physiological changes occurred because of this sleep-deprivation, affecting the body temperature, the food intake, the autonomic nervous system, and so on till death. The circadian cycle of course affects the sleep stages at the level of several structures: -​ the suprachiasmatic nucleus, -​ the pineal gland, -​ the brainstem centers, -​ the thalamocortical neurons for the thalamocortical synchronization and desynchronization, are fundamental to pass from an awake state to an asleep one and vice versa.  (1)​ The light plays the first key role in depolarizing specific ganglion cells in the retina containing the photopigment melanopsin. (2)​ These cells send their information to the anterior part of the thalamus in a specific nucleus known as suprachiasmatic nucleus of the thalamus, through the retinohypothalamic tract. The suprachiasmatic is thought to be the human master clock. (3)​ Then, through specific descending pathways, the information is sent to the preganglionic sympathetic neurons in the intermediolateral zone of the thoracic spinal cord, playing an important role in the vegetative response to light. (4)​ From the intermediolateral column, information is sent to the superior cervical ganglia (5)​ From the superior cervical ganglia sends the input to the pineal gland that synthetizes tryptophan as a precursor of the sleep-promoting hormone ‘melatonin’. The melatonin percentage in the human body is inversely proportional to the light in the surrounding environment, and an increased amount of this hormone in the blood affects the brainstem centers devoted to the awake-asleep passage. Superchiasmatic nucleus The superchiasmatic nucleus is considered as the “master clock.” Evidence for that lays in the fact that its removal in animals completely abolishes the circadian cycle rhythm of sleeping ad waking. This nucleus is quite important as it is located in the posterior part of the thalamus, since the hypothalamus can affect several aspects of both the vegetative and homeostatic balances (body temperature, hormone secretion such as cortisol, blood pressure, urine production, …). Stages of sleep EEG has been used to ‘stratify’ the different stages of sleep. First, there is the awake state characterized by the previously described features such as high frequencies and low amplitudes. Then, there are four non-REM phases and the additional REM phase itself. Passing from stage I to stage IV (deepest sleep stage) there is an evident progression leading to bigger amplitudes of the EEG signals, and a decrease in the detected frequencies. After more or less 60 minutes of sleep, the stages reverse occurs before entering the REM phase, characterized by features similar to the awake state. In the pictures below, several physiological parameters are recorded while changing during both a non-REM sleep and a REM sleep. In general, in the non-REM all the activities (including the muscle tone, the body movements, the heart rate, the penile erection, the breathing and metabolic activities) are reduced, reaching their lowest value at stage IV. -​ Slow, rolling eye movements -​ Decreases in muscke tone, body movements -​ Decreases is heart rate, breating, blood pressure, metabolic rate and temperature The opposite occurs during REM sleep. The penile erection is a fundamental marker to detect the REM phase and likely problems at the level of the reproductive system. -​ Rapid, ballistic eye movements, pupillary constriction -​ Paralysis of many large muscle groups -​ twitching of the smaller fingers, toes and the middle ear -​ increases in heart rate, BP and metabolism nearly as a high as in the awake state -​ spontaneous penile erection during REM sleep Movements during sleep REM such as the paralysis of many large muscle groups and the twitching of the smaller muscles in the fingers, toes, and the middle ear, verify as a consequence of a specific physiological chain of events. In fact, the relative physical paralysis during REM sleep arises from increased activity in GABAergic neurons in the pontine reticular formation, with the GABAergic being an inhibitory activity. This formation so projects to inhibitory neurons that synapse in turn with lower motor neurons in the ventral horn of the spinal cord, therefore the big muscles don’t move. Finally, there is an increased activity of descending inhibitory projections from the pons to the DCML dorsal column nuclei also causing a diminished response to somatic sensory stimuli. Reticular activating system Specifically talking about the role of the reticular formation and the interaction between thalamus and cortex within the whole neural circuit governing the sleep, in 1949, Horace Magoun and Giuseppe Moruzzi found that electrically stimulating a group of cholinergic neurons at the level of the reticular formation near the junction of the pons and midbrain causes a state of wakefulness and arousal, the reticular activating system. Their work was really important because implied that wakefulness was not just the presence of adequate sensory experience but required a special activating circuitry. At the same time a swiss physiologist, Walter Hess, found that stimulating the thalamus in an awake cat with low-frequency pulses produced a slow-wave sleep, usually associated to the deep phases of the sleep instead. These parallel experiments revealed the coordination mechanism between the reticular activating system and the thalamus at the base of the awake-asleep and asleep-awake passages. Eye movements – beginning of REM Coming to the origin of the rolling eyes event, it is important to investigate again the interactions among the thalamus, the reticular formation and the cortex. Endogenous signals from the pontine reticular formation are directly transmitted to the motor region of the superior colliculus which is involved in controlling and coordinating timing and direction of the eye movements. Although the EEG signal appears to be similar to the awake state during REM sleep, the neural origin is different: the REM sleep is characterized by EEG waves that originate from the activation of the pontine reticular formation by either an endogenous system or an electrical stimulation and then propagate through specific pathways, till the lateral geniculate nucleus of the thalamus and finally the visual cortical area. These pontine-geniculo-occipital (PGO) waves provide the neural substrate for the beginning of REM sleep. Regions activated REM and non-REM sleep The fMRI technique has been used in association to the EEG to precisely detect which regions were activated during REM and non-REM sleep. The results claimed -​ the dorsolateral prefrontal cortex being inactive during the REM sleep -​ lots of limbic areas and paralimbic areas, such as o​ the parahippocampal gyrus, o​ the pontine tegmentum and o​ the anterior and posterior cingulate cortices, being instead active during the REM sleep. This particular architecture, active during the REM phase, seems to be implicated in the mysterious phenomena of dreams. These higher centers can be modulated during the different stages, and for this purpose there are several nuclei providing and mediating the interaction among the reticular formation, the thalamus and the cortex. First, the cholinergic neurons of the reticular formation activate and project to the thalamus that in turn project to the cortex in a very diffuse and non-specific way. Usually, the activity of these neurons contributes both the awake phase and the REM stage. Then other nuclei, such as -​ the raphe nuclei located in the posterior part of the reticular formation immediately posterior to the cholinergic nuclei, hosting serotoninergic nuclei, -​ the locus coeruleus hosting noradrenergic neurons, -​ the tuberomammillary neurons of the hypothalamus hosting histamine-containing neurons coactivate and produce the awake state together. In the case this doesn’t occur, there must be a neurological problem. When taking histaminic drugs after an allergic reaction for example, they inhibit the histamine-containing neurons (tuberomammillary neurons), this way inducing a sense of tiredness. At the level of the ventrolateral part of the preoptic neurons there is a nucleus capable, by means of GABA, of inhibiting the activity of cholinergic, monoaminergic and histamine-containing neurons in order to regulate the asleep and awake phases. All the neurons seen so far for the activity in the brainstem possess the property of affecting the rhythmicity of the interaction between the thalamus and the cortex. The different sleep stages are affected and determined by the thalamocortical synchronization and desynchronization. In this regard, the thalamocortical neurons are present in two physiological states. -​ The first one is the tonic firing state and occurs when they are depolarized by the reticular activation system, and then transmit information to the cortex associated to the encoding of peripheral stimuli.  -​ The second physiological state is the oscillatory or bursting one, occurring when the thalamic and cortical neurons became synchronous and leading to a disconnection of the cortex from the world so that external stimuli cannot be perceived anymore (maximal disconnection). Still, I there is too much contact with the external world it could be difficult for this to verify and so to sleep well. The sleep spindle is another physiological marker for this phase and is mediated by the interaction between the thalamus and the cortex. Oscillatory or bursting states are stabilized by hyperpolarizing relevant thalamic nucleus by means of GABAergic neurons in the thalamic reticular formation. In the picture below on the right, thalamic neurons of the reticular formation are inhibiting the thalamocortical cells. But in general, excitatory and inhibitory connections between thalamocortical cells, pyramidal cells in the cortex, and thalamic reticular cells provide the basis for the sleep spindle generation. This mechanism ends up being called the thalamocortical feedback loop.​ In his book ‘The interpretation of dreams’, the famous neurologist Sigmund Freud divided the human phycological substrate subdivided into Ego, Es and Superego. His main idea was that during dreaming the Ego relaxed and stopped inhibiting the subconscious, finally letting dreams rising. More recently, according to scientists investigating both humans and other animals, dreaming evolved to dispose unwanted memories that accumulate during the day, but also to help in consolidating the learned tasks, perhaps by strengthening synaptic activity associated with recent experiences sleep-dependent improvement in learning. Q&A: Why do we need to sleep? Sleeping is important for maintaining the body in the right homeostatic range. Indeed, using animal models, scientists understood that, without sleeping, there would be several problems because in this phase there a lot of physiological changes associated to specific homeostatic parameters. However, what happened during the sleeping phase allowing human being to maintain the right equilibrium is not still understood.