Methods & Techniques in Neuroscience: Single Cell Recording

• Single-neuron recording offers the best temporal and spatial resolutions
• A diverse range of brain functions are represented by the activities at the micro (neuronal) level

But ...

• Only patients can be the participants
• Recording is very invasive and complex

Constraints and limitations:

• Poor generalizability, as only patients can be participants
• Total experiment time is limited (usually 1 week)
• Effect of medication
• Heterogeneity of participants (age, cognitive skills, task performance, etc)
• Brain areas of study are limited by clinical constraints
• Electrodes cannot be moved
• Number of brain areas to study simultaneously are limited
• Poorly understood spontaneous activity ("resting state")
• Possibility of microstimulation to the neurons to explore the causal role in perception

Multiple choice:

Place cells are found in:

• Primary visual cortex
• Orbitorfrontal cortex
• Hippocampus - yes
• Anterior cingulate cortex

The typical sampling frequency used for recording single neuron activity is:

• 2 kHz
• 20 kHz
• 200 kHz
• 200 Hz

Which of the following features is NOT used in spike sorting:

• Spike shape
• Spike amplitude
• Firing rate
• Spike's refractory period - yes

Which of the following techniques offers the best spatial resolution?

• fMRI - yes
• EEG
• TMS
• MEG

Choose the INCORRECT statement about action potential (AP:

• AP is all or nothing
• The profile of APs of one neuron can vary depending on the state of the neuron - yes; incorrect
• AP propogates in a non decremental fashion
• The duration of AP is 1ms

Short answer questions:

• List five ways to establish a link between single neurons and perception/cognition
• Describe five limitations of single unit recording in humans
• What is an Action Potential (AP) of a neuron?
• Briefly describe any four characteristics of AP
• What are the major differences between extrcellular and intracellular recording?
• Which recording is more suitable in human?
• What is spike sorting? List 3 features used in spike sorting
• What is the procedure to be followed in single-unit recording in humans

• Epileptic subjects: localizing seizure foci
• Depressive/OCD: Cingulotomy
• Parkinsonian
• Brain-computer-interface
• Neural prosthetics

Method:                 Type:                       Invasive?                 Signal:

TMS                     stimulation                  no               electromagnetic

EEG/ERP                 recording                    no                      electrical

MEG                       recording                    no                         magnetic

f-MRI                   recording                      no                     haemodynamic

PET                          recording                  yes                haemodynamic       

single cell                 recording                   yes                      electrical

Spatial resolution: The accuracy with which one measures where it is occurring

Temporal resolution: The accuracy with which one measures when it is occuring

• They are mostly correlational - patterns of brain activity during task - but no proof that the activity took part in the task - correlation does not imply causation
• No dissociation between excitatory and inhibitory activity
• More activation does not mean more processing!
• Spontaneous brain activity even in rest ie two conditions should be contrasted
• Null results are difficult to interpret - eg the area was active in both tasks - or, the areas was not active in either task, or, - too subtle differences to detect 

Glial cells

• More in number
• supporting neurons
• No direct role in information transmission?
• No role in electrical signalling

Nerve cells (Neuron)

• Basic building block
• More than 1000 varieties but with shared basic architecture
• Complexity of behaviour depends less on the specialisation of individual cells but more on the connectivity
• Generates electrical signals
• Communicates mainly via electro-chemical fashion

Four main regions:

• Cell Body - metabolic centre, contains nucleus, genes
• Dendrites - receiving signals from other neurons
• Axons - carrying signals to other neurons
• Synaptic Zone - connection zone between neurons

• Constitute the signals by which the brain receives, analyses, and conveys "information"
• Same basic form
• Information is conveyed, not by the form, but by the pathways
• 100 mV with a 1ms duration
• Propogated along the axon at a fixed and slow speed (m/sec)
• Nondecremental (same amplitude)
• All or none phenomenon (either it fires at its full amplitude, or it doesn't fire at all)

Extracellular Recording

• A small electrode (<10) is plunged blindly into the neural tissue and poked about until a single cell response is obtained
• Very high impedance
• Needs special pre-amplifier

Issues: Mixing of neurons/unit isolation, lack of anatomical details

Intracellular Recording

• Glass pipette electrode (<1) is inserted inside the cell membrane
• Pipette is filled with a solution with similar ionic composition to the intracellular fluid
• Ideal for recording graded potential leading up to spike discharge, or failing of spike discharge in a neuron
• Differentiates excitatory from inhibitory synaptic activity

Issues: cell damage, not very feasible in behaving animal

• Subjects are medical patients with pharmacologically resistant epilepsy
• Treatment - removal of brain tissues generating seizures
• Electrodes are surgically implanted to localise the focus of the seizures
• Mid temporal lobe (MTL) is the primary target area: hippocampus, amygdala, entorhinal cortex, parahippocampal cortex
• Average hospital stay = 7-14 days

Advantages of Single Unit Recording in Humans

• Signal source: direct recording
• spatial resolution: 1-2 mm (EEG - cm)
• Temporal resolution: ms
• signal-to-noise ratio: (less artefacts eg no blink artefacts)
• high frequency brain activity

Three conditions:

• they must have failed to respond to medical management (antiepileptic drugs)
• their quality of life would be markedly enhanced by achieving a seizure-free surgical outcome. The ideal surgical canditate is a young person whose educational and vocational opportunities  promise to be enhanced substantially by eliminating seizures.These patients also ar ideal subjects for invasive brain research.
• Third, their seizure focus must be localized to a circumscribed portion of the brain that can safely be removed surgically.

• No requirement of preoperative diagnostic testing
• Experiments inside the operating room
• Max duration of experiment = 30 min
• Noisy/challenging environment; no room for debugging faults
• Patient's responses are limited (simple verbalisation or button press)
• Patient is in supine position (minimal movement is allowed)

• There is some residual uncertainty about the location of seizure foci
• Requirement of additional diagnoistic testing (recording from electrode arrays chronically implanted)
• Experiments are carried out in a specially equipped hospital suite (less noisier than Acute recording)
• A lot of variability (devices, techniques, protocols) across epilepsy centers
• Patients are less restricted and more demanding tasks can be performed

Procedure: Chronic Recording

• Cognitive and sensory functions are tested through testing: Both pre- and post-operative testing are necessary in order to ensure that IQ, visual function, and other factors won't confound the data.
• Need careful documentation of the medical status: - Usually antileptic medication is lowered to increase chance of epilepsy - if no seizure is recorded, patient is often sleep deprived
• Testing is done several days after electrode implantation
• Testing window spans only for a few days when the patient is awake and alert, has minimal seizure, and not sleep-deprived
• Pre- and post- MRI scans are necesary to localize electrodes

See slide re Recording steps

Rate coding

• Information is in the mean firing rate of the neuron
• Most widely used and understood
• Robust for recognition and processing of static objects
• Less suitable for fast and dynamic action

Temporal coding

• Information is in the timing of the spike
• No precise definition, anything not based on rate can be based on timing (most important is neuronal synchronisation)

'Neural cocktail party problem'

Imagine a room where several people are sitting and talking. The likely range of people is known but the precise number of people is not. Their language is unknown to us and some of them talk at once. In this analogy, people correspond to NEURONS.

Imagine we have placed microphones around the room. Our task is to find out:

• How many people are talking
• Characteristics of the sounds eg pitch, loudness, variability, statistics
• what is the discourse of each person

Microphones - Electrodes

Loudness - spike amplitude

Pitch - spike waveform

Talkative - firing rate

• Align spike sequences with stimulus onset (or any event) which is repeated n times
• Divide the stimulus period S into N bins of size Z
• count the number of spikes k within individual bin for all trials
• Complete histogram

• Neurons should have discernible statistical differences in their firing patterns in response to the different external stimuli

• the firing patterns of the neurons should not be affected by the mere motor responses

• Fluctuations in the firing to the repeated presentation of identical stimuli sould be predictive of the observer' response

• Direct interference with the firing of neurons should lead to change in behavioural response

• Temporary or permanent removal of all or part of the candidate set of neurons should lead to measurable perceptual deficit

Responsive Neuron:

• Acitivity at task is different from activity at base line

(Category) Selective Neuron:

• Activity for the category is different from the baseline activity (t-test, Wilcoxon test)
• Differences in activity among categories (ANOVA)
• Difference in activity among pair-wise (Wilcoxon)
• No difference in activity to distinct individual stimuli within the category (ANOVA)

Examples on slides of:

'Famous face' neuron (hippocampus)

Single neuron and Object recognition

• Remarkable degree of segregation of categories at the level of single neurons
• Existence of sparse coding in the MTL
• Hippocampal neurons respond more to spatial objects
• It is possible to predict (to some extent) whether the preferred stimulus was presented or not!

Three types of neurons:

• visual selective
• imagery selective
• both selective

Tight coupling between perception and imagery

Role of MTL in imagery

See study on Visual Invariance (Quiroga, 2005) (Princess Di famous face study)

• Out of 132 neurons, 51 showed invariance to a particular, individual, landmark building, animal or object
• Are these neurons 'grandmother cells'?
• These cells make link between perception and long term memory
• Representation is fairly abstract and invariant
• Possible they are crucial for storing new associations

Think about a 2D model of emotion with valence on the x-axis and arousal on the y-axis.

Study: 'Emotion processing by PFC Neurons' (Kawasaki et al., 2001).

• Same neurons discharge during execution and perception of action
• So ... overlapping sensory-motor representation

Four types of reported neurons:

• action-observation neuron
• Action-execution neuron
• Action observation/execution non-matching neuron
• Action observation/execution matching neuron

Mukamel et al. (2010) study found:

• A significant proportion of neurons in Supplementary Motor Area (SMA), hippocampus responded to both observation and execution of actions
• A subset of neurons demonstrated excitation matching, another subset inhibition matching, and a final subset excitation-inhibition matching
• Existence of multiple systems with neural mirroring mechanisms for flexible integration and differentiation of perceptual and motor aspects of actions performed by self and others