Large-scale recording of neuronal ensembles Flashcards
Describe the two ways spike pattern variability has been treated
Spike threshold and pattern variability have been traditionally viewed as an indication of the brain’s imperfection, a noise that should be averaged out to reveal the brain’s true attitude toward the invariant input .
Alternatively, we may hypothesise that the ‘noise’, that is, the mismatch between the physical input and neuronal response, reflects self-organised patterns in the brain, and it is this centrally coordinated activity of cortical neurons that creates cognition
How can we obtain indications of the brain’s perspective of the environment?
Extracting the variant (brain-generated) features, including the temporal relations among neuronal assemblies and assembly members from the invariant features represented by the physical world might provide clues about the brain’s perspective on its environment.
What can we do with current technology and what is required to achieve the full potential of massively parallel neuronal recordings?
Currently, wire and micro- machined silicon electrode arrays can record from large numbers of neurons and monitor local neural circuits at work. Achieving the full potential of massively parallel neuronal recordings, however, will require further development of the neuron–electrode interface, automated and efficient spike- sorting algorithms for effective isolation and identification of single neurons, and new mathematical insights for the analysis of network properties.
How does Buzsaki use the metaphor of a orchestra to describe EEG/ MEG?
The first available method is to record the total noise generated by the orchestra but without the ability to distinguish the instruments and musicians. The dynamics of the continuous time-variable signal can be analysed by various mathematical means in the time and frequency domains, but these methods can reveal little about orchestration. This ‘temporally integrated field’ method is analogous to recording with electroencephalography (EEG) or magnetoencephalography (MEG) in the brain.
How does Buzsaki use the metaphor of a orchestra to describe fMRI/ PET?
A second method can take infrared pictures of the orchestra. This will measure the heat generated by the musicians’ muscle activity. Given the orderly arrangement of the instruments, the pictures taken during some pas- sages of the melody can identify spots of dominant activity, an approach analogous to functional magnetic resonance imaging (fMRI) or positron emission tomography (PET) snapshots taken from the living brain. Unfortunately, this ‘spatial mean field’ approach fails to capture the essence of music: temporal dynamics.
How does Buzsaki use the metaphor of a orchestra to describe single cell recordings?
A third method can sense the sound pressure generated by any one of the instruments and send a pulse to the observer whenever the pressure exceeds a certain threshold, analogous to recording of action potentials (spikes) emitted by single neurons in the brain. By monitoring different but single musical instruments of the same or even different orchestras over many successive performances and pooling the measurements as if they were recorded simultaneously, one can reconstruct some essential feature of the score
What does Buszaki comment regarding the viability of single cell recordings?
This independent ‘single-cell’ approach has yielded significant progress in neuroscience . However, this method would fail when applied to a jazz ensemble where the tune is created by the dynamic interactions among the musicians ‘on the fly’ and which interactions vary from performance to performance. It also largely fails when applied to central brain circuits where myriad ensembles are at work at multiple temporal and spatial scales.
What does Buszaki claim are the principal instruments in the arsenal of contemporary cognitive-behavioral neuroscience?
Field potential analysis, imaging of energy production in brain structures and single-cell recording techniques are the principal instruments in the arsenal of contemporary cognitive-behavioral neuroscience for the study of the intact brain.
What does Buszaki comment about the viability of this field potential analysis?
Even their combined, simultaneous application in behaving subjects falls short of the goal of explaining how a coalition of neuronal groups make sense of the world, generate ideas and goals, and create appropriate responses in a changing environment.
In the brain, specific behaviors emerge from the interaction of its constituents: neurons and neuronal pools. What does Buzsaki claim is required for studying these self-organised processes?
Studying these self-organised processes requires simultaneously monitoring the activity of large numbers of individual neurons in multiple brain areas. Recording from every neuron in the brain is an unreasonable goal. On the other hand, recording from statistically representative samples of identified neurons from several local areas while minimally interfering with brain activity is feasible with currently available and emerging technologies and indeed is a high-priority goal in systems neuroscience.
What other methods does Buzsaki claim can aid this task and to what extent should they play a role?
Many other methods, such as pharmacological manipulations, macroscopic and microscopic imaging and molecular biological tools, can aid this task, but in the end all these indirect observations should be translated back into a common currency—the format of neuronal spike trains—to understand the brain’s control of behaviour.
When a single electrode in placed near a number of neurons ad records from all of them, how can action potentials from different neurons be distinguished from the signal?
Because neurons of the same class generate identical action potentials (all first violins sound the same), the only way to identify a given neuron from extracellularly recorded spikes is to move the electrode tip closer to its body (<20 μm in cortex) than to any other neuron.
What is required to record from another neuron ‘with certainty’?
To record from another neuron with certainty, yet another electrode is needed.
Why does Buszaki claim that improved methods are needed for the simultaneous recording of neuronal populations?
Because electrical recording from neurons is invasive, monitoring from larger numbers of neurons inevitably increases tissue damage. Furthermore, understanding how the cooperative activity of different classes of neurons gives rise to collective ensemble behaviour requires their separation and identification. Because most anatomical wiring is local, the majority of neuronal interactions, and thus computation, occur in a small volume. In the neocortex, the ‘small volume’ corresponds to hypothetical cortical modules (for example, mini- and macro-columns, barrels, stripes, blobs), with mostly vertically organised layers of principal cells and numerous interneuron types. Thus, improved methods are needed for the simultaneous recording of closely spaced neuronal populations with minimal damage to the hard wiring.
What has ‘dramatically increased the yield of isolated neurons’?
The recent advent of localised, multi-site extracellular recording techniques has dramatically increased the yield of isolated neurons
How have the recent advent of localised, multi-site extracellular recording techniques increased the yield of isolated neurons?
With only one recording site, neurons that are the same distance from the tip provide signals of the same magnitude, making the isolation of single cells difficult. The use of two or more recording sites allows for the triangulation of distances because the amplitude of the recorded spike is a function of the distance between the neuron and the electrode.
What is the ideal layout of the electrode relative to the neurons?
Ideally, the tips are separated in three-dimensional space so that unequivocal triangulation is possible in a volume. This can be accomplished with four spaced wires (∼50 μm 18–20 spread; dubbed ‘tetrodes’).
What advantages do wired electrodes have over sharp tipped electrodes?
Wire tetrodes have numerous advantages over sharp-tip single electrodes, including larger yield of units, low-impedance (an expression of the opposition that an electronic component, circuit, or system offers to alternating and/or direct electric current) recording tips and mechanical stability. Because the recording tip need not be placed in the immediate vicinity of the neuron, long-term recordings in behaving animals are possible.
Why is it said that “there is a large gap between the numbers of routinely recorded and theoretically recordable neurons?”
A cylinder with a radius 140 μm contains ∼1,000 neurons in the rat cortex, which is the number of theoretically recordable cells by a single electrode. Yet, in practice, only a small fraction of the neurons can be reliably separated with currently available probes and spike sorting algorithms. The remaining neurons may be damaged by the blunt end of the closely spaced wires, or may be silent or too small in amplitude.
An ideal recording electrode has a very small volume, so that tissue injury is minimised. However, a very large number of recording sites is ideal for monitoring many neurons. Obviously, these competing requirements are difficult to satisfy. Describe two options for recording devices apart from wire electrodes and tetrode recording principles and what advantages they present
Micro-Electro-Mechanical System (MEMS)-based recording devices can reduce the technical limitations inherent in wire electrodes because with the same amount of tissue displacement, the number of monitoring sites can be substantially increased.
Whereas silicon probes have the advantages
of tetrode recording principles, they are substantially smaller in size. Furthermore, multiple sites can be arranged over a longer distance, thus allowing for the simultaneous recording of neuronal activity in the various cortical layers.
To what extent can current multi-shank probes record from multiple neurons?
Currently available muti-shank probes can record from as many as a hundred well-separated neurons
What does the geometrically precise distribution of the recording sites allow for?
Importantly, the geometrically precise distribution of the recording sites also allows for the determination of the spatial relationship of the isolated single neurons. This feature is a prerequisite for studying the spatiotemporal representation and transformation of inputs by neuronal ensembles.
What is the principle limitation of increasing the numbers of recording sites? What, however, is important to note about this?
The principal limitation of increasing the numbers of recording sites is the width of the interconnection between the recording tips and the extracranial connector (2 μm-wide connections with 2-μm space).
It should be noted, though, that industrial production presently uses 0.18 μm line features, and multiple levels of metal and much thinner interconnect lines are expected to become standard in coming years.
An indispensable step in spike-train analysis is the isolation of single neurons on the basis of extracellular features. Name the two broad classes of spike sorting methods and their assumptions
The first class attempts to separate spikes on the basis of amplitude and wave form variation on the assumption that neighbouring neurons generate invariant spike features.
The second general approach, triangulation, is based on the tacit assumption that the extracellularly recorded spikes emanate from point sources rather than from the complex geometry of neurons.
Evaluate the assumption of the first approach
The assumption that neighbouring neurons generate invariant spike features is difficult to justify in most cases.
Evaluate the assumption of the second approach
The tacit assumption that the extracellularly recorded spikes emanate from point sources rather than from the complex geometry of neurons is obviously a simplistic idea, because every part of the neuronal membrane is capable of generating action potentials. Because the extracellular spike is a summation of the integrated signals from both soma and large proximal dendrites , the extracellularly recorded spike parameters depend on the extent of spike backpropagation and on other state and behaviour-dependent changes of the membrane potential. These changes can affect the estimation of the neuron’s virtual ‘point source’ location and may place the same neuron at different locations, resulting in omission errors of unit isolation.
When is the spike amplitude variation most substantial?
The spike amplitude variation is most substantial during complex spike burst production, with as much as 80% amplitude reduction because of Na+ channel inactivation.
What do improved spike sorting methods incorporate?
Improved spike sorting methods therefore analyse not only the amplitude but also wave shape variation of spikes.
Describe another problem with the point-source assumption for action potentials
Another problem with the point-source assumption for action potentials is that the somatic origin is not always resolvable with distant recording sites. For example, in the rat neocortex, extracellular spikes can be recorded from the apical shaft of layer-5 pyramidal neurons as far 500 μm from the cell body. As a consequence, a single electrode tip , or a tetrode, placed in layer 4 can equally record from layer-4 cell bodies or apical dendrites of deeper neurons.
In the brains of what class of species is this classification error particularly apparent?
Such mis-classification errors are especially serious in the primate brain where spikes can be recorded from several hundred micrometers away from the cell bodies of large pyramidal cells.
How can the sources of unit-sorting errors be circumvented?
These sources of unit sorting errors can be circumvented by recording at multiple sites parallel with the axodendritic axis of the neurons. Importantly, such multiple-site monitoring can be exploited for the study of behaviour-dependent intracellular features and for resolving temporally superimposed spikes of different neurons
What is the major cause of unit isolation errors?
The amplitude and waveform variability of the extracellularly recorded spike is the major cause of unit isolation errors.
How do triangulation methods work?
Triangulation methods visually analyse two-dimensionally projected datasets at a time. With multiple site-recorded data, successive com- parisons of the various possible projections generate cumulative errors of human judgment. Cumulative human errors can be eliminated by automatic clustering methods of high-dimensional data.
What is a further difficulty of assessing spike sorting algorithms?
A further difficulty is that no independent criteria are available for the assessment of omission and commission errors of unit isolation. As a result, improvement of spike sorting algorithms is not guided by objective measures. In the absence of quantitative criteria for unit isolation quality, inter-laboratory comparison is difficult and makes interpretation controversial
What did a recent study result in regarding this problem of a lack of objective assessment tools for cluster cutting?
A recent study involving simultaneous intracellular and extracellular recording from the same pyramidal neurons resulted in a new database that allows for the objective assessment of spike classification errors, as well as the development of a semiautomatic clustering algorithm that is superior to manual clustering
After most or all instruments of the orchestra are isolated, the next step is the separation of strings from woodwinds and then oboes from clarinets. This is an important step because brain networks consist of several neuronal classes.
What is paradoxical about this in current neurophysiological practice?
Paradoxically, current neurophysiological practice rarely distinguishes among the various neuron classes. Unit classes are typically generated post hoc, as they relate to behaviour, without reference to the types of neurons that give rise to them.
However, the potential conclusion from an experiment reporting that all cortical pyramidal cells did one thing and all interneurons something else is qualitatively different from the conclusion that can be drawn from the information that 80% of all (unclassified) cells behaved differently from the rest. To understand the contribution of different neuron types to network activity, they have to be identified.
How can pyramidal cells be distinguished from interneurons without imaging tools or direct observation in certain parts of the brain?
In the hippocampus, several features, such as spike duration, firing rate and pattern, spike waveform and the relationship to field patterns, can be used to separate pyramidal cells from interneurons and some interneuron classes from each other. Similar classification criteria are not yet available in the neocortex.
What could facilitate this classification process in the neocortex?
Large-scale, high-density recording of neurons can facilitate the classification process. This is because a small percentage of cell pairs show robust, short-timescale correlations, indicative of monosynaptic connections. Monosynaptic excitation and inhibition can
thus identify excitatory and inhibitory neuronal classes, respectively. In turn, the extracellular features of the identified minority can be used as a template for classifying the remaining population.
A necessary requirement for understanding the transformation of inputs by a neuron or neuronal assemblies is information about both their input and output. How do our current capabilities fall short with this?
In contrast to recording spike output with high precision, no method is available for monitoring all inputs at the resolution of dendrites and spines of single neurons.
What are two potential strategies for recording these inputs?
Large-scale monitoring of spike patterns of presynaptic cell assemblies to a single neuron is a potential strategy.
An alternative compromise is measuring the extracellular current flow that reflects mainly the linearly summed postsynaptic potentials from local cell groups. Recording the voltage gradients by geometrically arranged sites of silicon probes allows current densities to be calculated for an estimation of the mean input to the neuron group in the recorded volume. High density recording with silicon probes therefore can be used not only to monitor the spike output from neuronal groups, but also to estimate their summed input. High spatial resolution of extracellular currents together with spiking activity of neurons from which these currents emanate provides a means for assessing input–output relations of neuronal coalitions.
