Notes on Neurons & Neural Firing

Neurons & Neural Firing — Study Notes

Neurons: main parts and their basic roles

• Neurons are not physically fused to one another; they connect and communicate via synapses.
• Neurons come in many types, differing in shape, size, connections, and function.
• The three main parts of a typical neuron:
  • Dendrites: receive signals from other neurons or sensory receptors; input structures with branching to collect information.
  • Soma (cell body): metabolic center of the neuron; integrates incoming signals from dendrites.
  • Axon: long projection that conducts electrical impulses away from the soma to axon terminals; ends in synapses.
• Dendrites and axons are both processes extending from the soma, but their roles differ: dendrites primarily input, the axon output travels the signal to other neurons.
• In sensory neurons, receptor cells act as input sites for physical signals; interneurons are a class of neurons involved in processing within the CNS.

How information is transferred in the nervous system

• Signals can be received in chemical form (neurotransmitters) or physical form (sensory signals).
• Transduction: Sodium diffuses down through the dendrites and the cell body, contributing to the generation of an electrical signal.
• The electrical impulse travels quickly along the axon to the axon terminals.
• At the synapse, neurotransmitters are released, triggering the process anew in the next neuron.
• This transmission involves synaptic communication rather than direct physical connections between neurons.
• The classic question guiding investigation: How is information transferred in the nervous system? And how is information represented in various cognitive functions (perception, attention, memory, language, decision-making, problem solving, consciousness)?

How information is represented in the nervous system

• Representation of cognitions is linked to patterns of neural activity across populations of neurons.
• Key cognitive domains under study include perception, attention, memory, language, decision-making, problem solving, and consciousness.
• Philosophical context: Realism and the neural basis of perception—how brain activity interprets electrical signals as experiences (reference to the Morpheus quote about reality).

Action potentials: basics of neural signaling

• Action potentials are electrical signals that travel along axons; crucial properties:
  • Occur only in axons.
  • Can travel at speeds up to the order of
    $v_{AP} \,\approx\, 100\ \text{m/s}.$
  • Amplitude (strength) remains constant as the spike propagates along the axon (all-or-none transmission).
  • Propagation is reliable without loss of signal strength along the axon.
• Resting potential:
  • The resting potential is typically between $V_{rest} \,=-60\text{ to }-80\ \text{mV}.$ (intracellular relative to extracellular space)

Steps in an action potential (classic sequence)

1) Sodium channels open in a portion of the axon, initiating depolarization.
2) The depolarization spreads; if threshold is reached, the action potential is triggered and propagates along the axon.
3) As depolarization proceeds, potassium channels open, leading to repolarization.
4) The ion pumps/restoration mechanisms reset the normal distribution of Na+ and K+ ions (Na+/K+ pump) to re-establish the resting state.
5) After the spike, a refractory period follows during which the neuron is less (absolute refractory) or more (relative refractory) resistant to firing another action potential.

• Note: A local region of the axon can begin to depolarize neighboring regions, causing a wave of depolarization along the axon.

Action potential details: ion channels and wave propagation

• Sodium channels open to allow Na+ influx, depolarizing the membrane (positive feedback loop).

• Sodium influx initiates depolarization; if depolarization reaches threshold, the entire axonal segment fires.

• Potassium channels open to allow K+ efflux, contributing to repolarization and eventually hyperpolarization.

• The ion pump restores the original distributions of Na+ and K+ ions after the spike.

• Refractory period ensures one-way propagation and regulates firing rate.

• Myelin and conduction:

  • Myelin wraps around some axons, reducing ion channel availability along the myelinated segments.
  • Conduction occurs via saltatory conduction: the action potential effectively jumps from node to node, increasing speed and efficiency of transmission.
  • Between nodes of Ranvier, ion channels are sparse or absent, so the spike effectively travels by jumping, rather than continuously.

All-or-None Law and its implications for cognition

• All-or-None Law: a neuron either fires a full action potential or it does not; the amplitude does not vary with stimulus strength.
• Implication: variations in perceived intensity and quality of stimuli are encoded not by stronger spikes in a single neuron but by:
  • Firing rate (rate coding) of individual neurons.
  • Recruitment of additional neurons (population coding).
  • Temporal patterns across neurons (temporal coding).
  • Combinatorial patterns across neural circuits (distributed/ensemble coding).
• Therefore, the strength of a stimulus is represented by neural activity patterns rather than increased spike amplitude.

Coding of stimulus intensity (intensity coding)

• Intensity coding asks: How strong is a stimulus?
• How is a more intense pressure stimulus represented neurally?
  • Higher firing rates in responsive neurons for stronger stimuli.
  • Greater recruitment of additional neurons within a population (population coding).
• The exact coding strategy may involve a combination of rate coding and population coding depending on the neural system.

Quality coding: how the brain encodes 'what' a stimulus is

• Question: What type of stimulus is it?
• Possible coding schemes for stimulus quality include:
  • Temporal Codes: information carried by the timing of spikes (e.g., sequences like tap, tap, tap, tap versus tap, tap, …, tap, tap).
  • Specificity Coding, Sparse Coding, and Population Coding (three proposed schemes for representing qualities).
• Visual and other sensory systems may rely on specialized patterns of activity to distinguish different stimulus qualities.
• Figures in lectures often show mapping of images to coding types; students are asked to identify which coding type best fits each example.

Specific coding schemes for quality (A, B, C)

• A. Specificity Coding: a neuron or small set responds to a highly specific stimulus (high selectivity).
• B. Sparse Coding: a relatively small subset of neurons is active for a given stimulus, providing efficiency and distinct representations.
• C. Population Coding: many neurons participate; information is distributed across a broader network.
• Question prompts in the lecture: Which image best represents each coding type? Why is this likely for coding quality?

Temporal coding and feature detectors in the visual system

• Temporal coding: information can be carried by the precise timing of spikes (e.g., rhythm or sequence of taps).
• Feature detectors: specialized neurons respond to specific features of stimuli, particularly in the visual system (e.g., edges, orientations, motion).
• Experience-dependent plasticity: the brain’s wiring adapts based on experience, refining feature detectors and perceptual representations.
• Related resources mentioned in class include brief videos on feature detectors and plasticity (e.g., YouTube links cited in the lecture).

Progressive complexity of representations in the temporal cortex

• Neurons further along the temporal cortex respond to increasingly complex stimuli.
• Conceptual question: Could a single neuron eventually code for a highly complex, highly specific stimulus (e.g., a famous face)? This touches on the idea of specialized “grandmother cells” and their limitations.
• The Steve Carell neuron (Quiroga et al., 2008) is cited as a vivid example of highly selective coding: a neuron that responds selectively to a specific face (Steve Carell).
• The slide also poses a provocative question about whether, with enough specialization, a neuron could represent a complex, recognizable entity (e.g., a celebrity like Michelle Obama), illustrating debates about neural coding granularity.

The Steve Carell neuron and implications for coding specificity

• The Quiroga et al. study is a prominent example of highly selective neural coding, where a neuron responds preferentially to a specific individual’s face.
• This finding informs discussions of grandmother-cell hypotheses and distributed vs. localist representations.
• It also motivates consideration of how recognition emerges from networks of neurons rather than single neurons acting in isolation.

Connections to cognition, philosophy, and real-world relevance

• The lecture frames cognition through neural substrates: perception, attention, memory, language, decision-making, problem solving, and consciousness.
• Philosophical context: Understanding what counts as 'real' involves interpreting electrical signals as meaningful experiences through neural processing, echoing themes from The Matrix and discussions of brain–mind relationships.
• Practical implications include designing experiments to probe neural substrates of cognition, interpreting neuroimaging data, and understanding how embodiment and sensory signals contribute to conscious experience.

Summary connections and practical takeaways

• Neurons have distinct input and output regions (dendrites vs. axons) and rely on synaptic transmission to communicate.
• Information transfer begins with sensory or chemical signals, followed by electrical signaling (action potentials) and chemical signaling (neurotransmitter release) at synapses.
• Action potentials have a stereotyped, all-or-none nature, with a typical resting potential of $V<em>{rest} \in [-80, -60] \text{mV}$ and conduction speeds up to $v</em>{AP} \approx 100 \text{m/s}$.
• The propagation of signals is facilitated by myelin (saltatory conduction) and is constrained by refractory periods to regulate firing.
• Stimulus intensity is represented by firing rate and population activity rather than by increasing spike amplitude.
• Stimulus quality is coded through multiple schemes (specificity, sparse, and population coding) and temporal coding strategies.
• Feature detectors and experience-dependent plasticity illustrate how neural representations become progressively specialized, contributing to conscious perception and recognition of complex stimuli.
• Concrete neural evidence (e.g., Steve Carell neuron) underscores the existence of highly selective neural representations, while also motivating discussion of network-based versus single-neuron coding.