Posted comment on ´Theta rhythmic clock-like activity of single units in the mouse hippocampus` by L. Zhang, X. Ma, G. Chen, E. Barkai and L. Lin and published in Journal of Neuroscience 36(16) 2016 p.4415 doi.org/101523/JNEUROSCI.3851-15.2016
In their article, Zhang and colleagues describe their finding of a small group of neurons in the mouse hippocampus that exhibit theta oscillations only during waking exploration and REM sleep. The cells were named theta clock-spiking cells and the theta oscillations exhibited by these cells differed to those oscillations normally found in the hippocampus as part of the information processing and memory functions for example.
Zhang and colleagues took 15 C57/6J freely behaving mice and used drivable microelectrode arrays to record the brain wave oscillations. Three 36-pin connector arrays were positioned in parallel and for the recording microdrive two independently movable electrode bundles of 12 tetrodes and 24 stereotrodes were implanted into both sides of the dorsal hippocampi. The connector pin arrays were connected to amplifiers and the extracellular signals were filtered through these amplifiers to separate neuronal activity and local field potentials (LFP). Spike signals and the LFP signals were filtered on-line at different frequencies and spike waveforms, time stamps, and LFP signals were saved to Plexon data files. Individual neurons were identified and sorted by clustering methods using Offline Sorter version 2.8 software. Interneurons were identified by their greater than 5Hz firing rates and 100-300 μs spikes and interspike intervals (ISIs) were used to further isolate theta clock-spiking cells from the other cells. Oscillations were identified using bandpass filtering with theta oscillations LFP at 4–12 Hz, gamma 30–80 Hz and ripple oscillations 100–250 Hz. Stationary data was selected for further analysis using an augmented Dickey–Fuller test and power spectral density analyses were performed on both spike and LFP signals. Theta phase locking firing was analysed using a Hilbert transform to split into instantaneous amplitude and phase components, followed by spike phase and phase locking calculations. Animal behaviours during the testing process were digitally recorded and positions and firing rate maps constructed. Locomotion velocity and firing rates were also calculated.
Their investigation led Zhang and colleagues to identify a small subset of neurons (5 cells) in the stratum oriens region of the hippocampal CA1. The cells were named theta clock-spiking cells and they were found to only persistently fire during REM sleep and in the theta states of waking exploration. They were silent during slow-wave sleep. The theta oscillations exhibited by these theta clock-spiking cells had a simple clock-like spike firing pattern with one spike per theta cycle. There was a significant difference between the average firing rates of these cells and the peak frequencies of power spectra of corresponding LFP theta neurons and therefore, it was concluded that from the five cells, four bore no relationship to theta LFP. The authors calculated firing rates of the theta clock-spiking cells under different theta states and found that the average firing rate during waking exploration was about 9Hz which was slightly higher than the peak frequency of the power spectrum of the corresponding LFP theta rhythm in the hippocampus at 8Hz. The average firing rate during REM sleep was also found to be different for the 4 cells with the theta clock-spiking cells exhibiting a theta frequency of about 5Hz with other cells with the LFP theta rhythm having a slightly higher frequency of about 7Hz.
Zhang and colleagues also looked at other cells in the hippocampus CA1 area. Out of 508 cells, they found 44 theta locked interneurons, 30 theta unlocked interneurons in the stratum oriens and stratum pyramidale and 425 complex spiking cells in the stratum pyramidale. Therefore, their subset of theta clock-spiking cells was very small. The method of cell separation meant that these cells were together where neurons were sparsely distributed. The authors also found ripple oscillations with these cells, but these were smaller than for cells of the stratum pyrimadale area and the theta oscillations had a delayed 3–24 ms phase, which indicated that the soma of the theta clock-spiking cells were probably located in the stratum oriens. The cells also showed a different peak interval of ISI distributions to the other theta oscillating cells of the area. A difference was found between the firing rates of the theta clock-spiking cells and the complex firing cells observed during wakeful exploration and also by differences in spike durations between the theta clock-spiking cells and interneurons.
An investigation into gamma oscillations led the authors to the conclusion that the theta clock-spiking cells were not locked to gamma oscillations in REM sleep nor wakeful exploration, or to ripple oscillations in SWS. The cells, although they showed non spatial preference, appeared however to be linked to locomotion velocity.
Therefore, Zhang and colleagues found in their study a very small subset of theta oscillating cells in the hippocampus CA1 which exhibited firing under two conditions; REM sleep and waking exploration. They hypothesized that these theta clock-spiking cells may provide a temporal reference in theta-related temporal coding or decoding of information in the hippocampal area, but unlike the place cells of the area, they did not encode spatial information. Owing to the correlation between the theta clock-spiking cells firing and locomotion velocity the authors hypothesized that there may be link between this small subset of cells and speed.
What makes this article interesting is firstly, the way in which a small population of cells can be detected and secondly, the complexity, both temporally and frequency-related, of brain waves. Accurate detection of small cell populations could lead to more precise attributes of function to brain area and could also lead to experimental and therapeutic methods where manipulation of small populations of cells only could lead to widespread functional and structural effects. Understanding brain waves and their functions in smaller cell populations could provide a means of manipulation, eg. by specific transcranial magnetism, that could result in widespread neuronal effects. Although, the main emphasis in Zhang and colleagues` study was the theta oscillation, their hypothesis could apply to all of the brain wave types.
Unfortunately, one of the problems with Zhang and colleagues experiments was the low sample number of theta clock-spiking cells found – only 5 cells out of 15 mice out of over 500 cells looked at and even one of the 5 was not the theta clock-spiking cell in question. Several reasons can be brought forward to explain such a low number of cells in an area known for its cognitive function. For example, the theta clock-spiking cells could represent anomalous readings. However, this is probably unlikely since theta oscillations were definitely different in the theta clock-spiking cell to the theta oscillations observed with other cells and those differences were observed over different situations, eg REM sleep and waking exploration. Support for such cells also comes from other species since clock spiking cells have also been reported in the optic lobes of insects as early as 1965. The small subset of cells identified by Zhang and colleagues could also represent immature cells or cells not at the same point of their life cycle relative to other theta oscillating cells in the area. This possibility is also unlikely since all theta clock-spiking cells were found in one area only, the stratum oriens, and the cells were not morphologically different to the other hippocampal cells. Another explanation is that the results represent cells active in a common behaviour which is not displayed by this specific mouse strain. This explanation could be considered possible since the hippocampus exhibits neurogenesis and cell function adapts to cope with the animal`s behavioural requirements. If this mouse strain is not very exploratory for example, maybe the number of cells responsible for this function is decreased relative to other mouse strains and therefore, a low number of cells would be observed. Similarly, the results could represent cells active with a very specific function. This is another possible explanation although in this case this specific function must be in low demand in this mouse strain or in the day-to-day life of these mice.
Therefore, since we assume that the small population of theta oscillating cells identified by Zhang and colleagues is functionally and detectably different to the other theta oscillating cells of the hippocampus it is necessary to determine why they are there. Zhang and colleagues investigated whether the firing pattern of these theta clock-spiking cells could contribute to the hippocampal self-generated theta oscillations in general since several intrinsic, atropine-resistant (ie. not cholinergic cells) theta generators have been found in the CA1 using isolated rat hippocampal preparations. However, since the author`s experiments showed that the theta clock-spiking cells demonstrated a different frequency of firing rate from the peak frequency of the theta oscillating cells contributing to the local field potential, their involvement in cognitive functioning such as information processing and memory from a content point of view was deemed unlikely. However, Zhang and colleagues did speculate that the activity of the theta clock-spiking cells may provide an overall time reference for the theta phase precession of intracellular membrane potential oscillations in place cells. Therefore, the theta clock spiking cells could play a temporal role. The authors also did find a specific function linked to the small subset of cells since there were correlations between the firing of the cells and the locomotive velocity of the animals during waking exploration. It was concluded then that there could be a link between theta oscillations and speed and this has been reported in other studies as well.
So, how can we explain theta oscillations, locomotive speed and the hippocampus? In this case of being awake and exhibiting exploratory behaviour, there are two sources of inputted information: visual speed (ie. the change in visual information inputted of the environment while the animal is exploring) and ´run` speed (ie. the speed of mouse movement). Previous research has shown that input of both occurs via the V1 visual cortex, with the input and interpretation of the information involving the hippocampus. The hippocampus is known to be strongly correlated to cognitive functions such as memory and spatial navigation, both required in exploratory behaviour and there is a link between sensory input (visual information in visual cortex) and object and location of spatial information in the hippocampal place cells during waking exploration. This link is not only demonstrated at the cellular mechanical level, but also through brain waves. Brain waves represent synchronous firing of cells and the frequency of the brain wave demonstrates the speed of neuronal firing at that time. By measuring the brain waves of one area or between areas, functioning and connectivity can be observed. For example, firing between the thalamus and cortex is activated in a specific temporal sequence and this connectivity can be modulated by inhibiting the inputs from the thalamus reticular nucleus which is GABA dependent. Another example involves the prefrontal cortex which is also important in memory recall and is linked to increased theta oscillations in temporal order maintenance whereas alpha oscillations are required for item maintenance. The hippocampal theta bursts drive the generation of prefrontal cortical theta-gamma dependent hippocampus coupling and firing of the enterorhinal cortex. Theta oscillations are also linked to memory and in the case of waking exploration, the mouse uses its spatial memory for interpretation as well as storage of information for future use. Formation of new memories involving the CA1 neurons occurs with encoding at pyramidal cells preferentially timed later than the theta oscillation peak coincident with input from the enterorhinal cortex and retrieval of memories occurring at the theta oscillatory trough coincident with firing input from the CA3 region. Lesions of the enterorhinal cortex lead to disruption of these theta oscillations and silencing of the CA3 neurons resulting in loss of temporal coding in the CA1. However, the authors demonstrated in their experiments that the theta oscillations observed from their theta clock-spiking cells were different to those of the pyramidal cells and therefore, it is unlikely that these cells are directly responsible for information processing and memory formation of the event. However, it is possible that the theta- clock spiking cells although not directly responsible for the content of the event during the active times of the waking exploration as is the normal function of the theta oscillations, provide instead a ´background pulse` for times of intervals in the exploratory behaviour, ie. akin to a drum beating time. Spontaneous firings of the 4 cells would keep the area in a state of ´readiness` whilst active place cells undergo the biochemical refractory periods necessary during continual firing periods. This is seen with saccades in retinal cells and incoming visual information. Refractory periods of the active visual cells means that priority of event characteristics is shifted to the unattended features and there is temporary inhibition of return so that the cells representing the important event characteristics can biochemically recover ready for the next wave of firing.
This explanation could also apply to the other scenario where theta clock-spiking cells are observed, that is in REM sleep. In this case, the mouse undergoes no exploratory behaviour, but is motionless with no visual input and therefore, functioning of this particular sub-group of cells cannot be contributed to visual speed and run speed, or place cell activity recording object and location. However, just like in the waking exploration scenario, in REM sleep there is predominately another brain wave frequency representing informational content and manipulation. In this case, the frequency of the brain wave activity is beta with interspersed low frequency theta oscillations. The function of the brain waves here is just like in the waking case, to represent synchronous firing of groups of cells, but in the case of REM sleep, the firing is linked to memory formation and consolidation. This function is supported by the observation that REM sleep is disrupted by inducing sleep apnoea and this leads to significant negative effects on spatial navigational memory. Therefore, what function could Zhang and colleagues population of theta clock-spiking cells have in REM sleep? Just like in waking exploration, this subset of cells could be the ´default` cell, providing the ´background pulse`, essentially active when the firing cells representing the event features during this memory formation period reach their refractory periods. In REM sleep, the frequency of the normal oscillatory rhythm for memory formation and consolidation is also beta with spiking theta rhythms. This combination of primary frequency and secondary frequency can also be observed under other circumstances. In NREM stage 2 sleep there are sleep spindles observed with theta oscillating cells as spikes, or in slow wave sleep there are bursts of sleep spindles where new information is being integrated, replay is seen and there is reconsolidation of memories. Therefore, like a temporal marker, the theta clock-spiking cells in REM sleep could be ´innate` markers spontaneously firing to maintain area ´readiness` whilst other cells biochemically recover from firing. This hypothesis is supported by the observation by Bernardi that sleep deprivation, known to be linked to poor memory recall, leads to region specific increases in theta oscillations suggesting that theta oscillations represent transient neuronal states unrelated to event content.
Therefore, this article is interesting because it demonstrates just how complicated neuronal firing is and how monitoring of brain wave functioning has to be carried out on much smaller scales if we are to determine how cell firing is linked to information processing and memory. It could be that the theta clock-spiking cells identified by Zhang and colleagues are just ´artifacts` or a spurious observation of a few hippocampal cells, but they could be the ´default` firing cells of this important area keeping it in ´readiness` whilst other cells biochemically recover. Input and binding of information and interplay between the hippocampus, enterorhinal cortex and cortical areas may focus research attention on the predominant brain waves and cell firing during event characteristic registration, but if the theta clock-spiking cells are linked to the ´default` state of the hippocampus then disruption of their functioning may prevent correct informational input and interpretation overall. For example, Alzheimer disease is linked to hyperexcitability of the hippocampal region and there are currently no hypotheses on how this comes about. It could be that the fault lies not with the neuronal cells involved with informational input, but with cells like the theta clock-spiking cells who are involved in the correcting functioning of the area, but who are not event related. Therefore, investigation of small groups of cells is important.
Since we`re talking about the topic……………………………….
…..can we assume that if the theta clock-spiking cells are linked to a specific activity which is not in much demand during waking exploration, if the mice were trained to perform a task dependent on mouse speed, then we should see a major increase in number of this subset of cells if their function is truly linked to locomotive speed?
…… since cannabinoids disrupt theta oscillations in the hippocampus and ketamine increases theta oscillations in the medial frontal cortex, if Zhang and colleagues` experiments were performed again with these drugs pre-administered would we see how the theta clock-spiking cells are linked to normal brain wave functioning in these areas?
……can we assume that if the mouse is exposed to anaesthetics and brain waves are monitored we would see more and more of the brain going into slow wave oscillations as normal, but we would see an effect on the theta clock-spiking group of cells?
…….sleep deprivation leads to region specific homeostatic increases in theta oscillatory activity and therefore, would there be an increased number of theta clock-spiking cells if these cells are linked to the ´default` firing state of the hippocampus?
…….is it possible that the GABA agonist, zolpidem, which leads to increased sleep spindles and increased REM would produce noticeable effects on the number of the theta clock-spiking cells?