Fundamental Glossary of Computational Neuroscience

How to Use This Glossary

• Bolded term → concise definition ≤ 35 words.
• ↪ related shows key cross‑links to other terms.
• Simple math/logic expressions appear in inline code (tau = R × C).

Tip: Treat these entries like function docs—each gives purpose, inputs, outputs, and where it fits in the neural codebase.

---

A–C

• Action Potential (AP): Rapid (~1 ms) depolarization–repolarization spike enabling long‑distance axonal signaling; threshold‑crossing follows all‑or‑none law. ↪ Hodgkin–Huxley, Ion Channel, Refractory Period
• Astrocyte: Glial cell regulating extracellular ions, neurotransmitter uptake, and metabolic support; forms tripartite synapse. ↪ EAAT, Gliotransmission
• Attractor Network: Dynamical system whose state space contains stable points/manifolds representing memories or decisions (e.g., Hopfield net). ↪ Recurrent Network, Energy Landscape
• Backpropagation Through Time (BPTT): Gradient‑descent learning for recurrent nets, unfolding temporal steps; biologically debated. ↪ STDP, Eligibility Trace
• Bayesian Inference: Probabilistic framework where posterior ∝ likelihood×prior, hypothesized mechanism for cortical processing. ↪ Predictive Coding
• Burst Firing: Groups of ≥ 2 spikes with ISI < 5 ms, enhancing synaptic transmission reliability and neuromodulator release. ↪ Rate Code, Temporal Code
• Calcium Spike: Dendritic or somatic depolarization mainly via Ca²⁺ channels, triggering plasticity, hormone release. ↪ NMDA Receptor, Plateau Potential
• Coincidence Detector: Neural mechanism responding maximally when inputs arrive within tight temporal window (e.g., NMDA receptor, ITD neurons). ↪ Temporal Summation
• Computational Map: Topographic representation transforming spatial variables (e.g., retinotopy, tonotopy). ↪ Somatotopy, Place Cell
• Convolution: Weighted sliding dot‑product; models receptive‑field filtering in vision. ↪ Gabor Filter, CNN

D–F

• Deep Brain Stimulation (DBS): Implanted electrodes delivering patterned currents to modulate circuit dynamics in Parkinson’s, OCD. ↪ Closed‑Loop Neuromodulation
• Dendritic Spike: Local nonlinear events in dendrite producing branch‑specific amplification. ↪ Backpropagating AP, Compartmental Model
• Diffusion MRI (dMRI): Imaging modality tracking water diffusion to infer white‑matter tracts. ↪ Connectome
• Dynamic Causal Modeling (DCM): Bayesian framework estimating directed connectivity from fMRI/EEG. ↪ Effective Connectivity
• E‑I Balance: Condition where excitatory and inhibitory currents roughly cancel on average, stabilizing network activity. ↪ Homeostasis, Criticality
• Eligibility Trace: Transient memory of synaptic activity enabling delayed reinforcement signals. λ controls decay. ↪ Three‑Factor Learning
• Ensemble Coding: Population‑level representation where information distributed across many neurons. ↪ Population Vector
• Excitotoxicity: Neuronal death from excessive glutamate‑driven Ca²⁺ influx. ↪ EAAT, Stroke
• Fano Factor: Variance/mean ratio of spike counts; ≈1 for Poisson, <1 regular, >1 bursty. ↪ Noise Correlation
• Firing Rate: Spikes per second averaged over window; canonical descriptive variable. ↪ PSTH

G–I

• GABA: Primary inhibitory neurotransmitter binding GABA_A/B receptors; counterbalances glutamate. ↪ IPSP, Benzodiazepine
• Gain Modulation: Scaling of neuronal response by multiplicative factor (e.g., attention, neuromodulators). ↪ Divisive Normalization
• Gliotransmission: Release of transmitters (ATP, D‑serine) from glia modulating synapses. ↪ Astrocyte
• Glutamate: Main excitatory neurotransmitter; see dedicated handbook. ↪ AMPA, NMDA, EAAT
• Hebbian Plasticity: “Cells that fire together wire together”; ∆w ∝ x · y. ↪ STDP, LTP
• Hodgkin–Huxley Model: Four‑ODE circuit describing squid axon AP; parameters = g_Na, g_K, g_L, C_m. ↪ Compartmental Model
• Homeostatic Plasticity: Slow adjustments (synaptic scaling, excitability) maintaining firing set‑points. ↪ E‑I Balance
• Information Rate: Mutual information per time between stimulus and spikes; bounds coding efficiency. ↪ Shannon Entropy
• Interneuron: Typically GABAergic local circuit neuron modulating timing, gain. ↪ PV Cell, SOM Cell
• Intrinsic Plasticity: Activity‑dependent change in membrane properties (ion channel density). ↪ Homeostasis

J–L

• Jitter: Variability in spike timing across trials; lowered by synchrony. ↪ Reliability
• Kalman Filter: Recursive estimator for linear‑Gaussian systems; analog to cerebellar error correction. ↪ State Space Model
• Kullback–Leibler Divergence (KL): Measure of distribution mismatch; used in free‑energy principle. ↪ Variational Inference
• Lateral Inhibition: Circuit motif sharpening contrast via inhibitory surround. ↪ Difference‑of‑Gaussians
• Long‑Term Depression (LTD): Lasting synaptic weakening (minutes–hours); often Ca²⁺‑low, mGluR‑dependent. ↪ LTP, STDP
• Long‑Term Potentiation (LTP): Enduring synaptic strengthening; canonical memory substrate. ↪ Hebbian, NMDA
• LFP (Local Field Potential): Low‑frequency (<300 Hz) extracellular signal reflecting population synaptic currents. ↪ EEG, CSD

M–O

• Markov Chain Monte Carlo (MCMC): Sampling method for posterior approximations; theoretical link to neural sampling. ↪ Gibbs Sampling
• Metaplasticity: Plasticity of plasticity rules (e.g., BCM sliding threshold). ↪ Homeostatic Plasticity
• mGluR (Metabotropic Glutamate Receptor): G‑protein‑coupled receptor modulating excitability. ↪ Glutamate
• Multiplexing: Encoding multiple variables in distinct dimensions (rate, phase). ↪ Phase Coding
• Neurotransmitter Release Probability (Pr): Chance vesicle fuses per AP; key for short‑term plasticity. ↪ Paired‑Pulse Ratio
• Noise Correlation: Shared trial‑to‑trial variability between neurons; constrains population code. ↪ Fano Factor
• Optogenetics: Genetically encoded light‑gated ion channels enabling circuit‑specific control. ↪ Channelrhodopsin, Photostimulation
• Oscillation: Rhythmic population activity (delta–gamma bands) coordinating communication. ↪ Phase Coupling

P–R

• Place Cell: Hippocampal neuron firing in specific spatial loci; supports cognitive map. ↪ Grid Cell
• Predictive Coding: Hierarchical model minimizing prediction error between top‑down and bottom‑up signals. ↪ Bayesian Inference
• PSP (Postsynaptic Potential): Voltage change from synaptic input; EPSP (depolarizing) vs IPSP (hyperpolarizing). ↪ Quantal Content
• Population Vector: Weighted sum of preferred‑direction vectors predicting movement. ↪ Ensemble Coding
• Quantal Size: Voltage change per vesicle release; sets synaptic gain. ↪ PSP
• Receptive Field (RF): Stimulus region modulating neuron firing; computed via reverse correlation. ↪ Convolution
• Refractory Period: Time post‑spike where AP generation suppressed; absolute vs relative. ↪ Hodgkin–Huxley
• Reward Prediction Error (RPE): Delta between expected and received reward signaled by dopamine bursts. ↪ TD Learning

S–U

• Shannon Entropy: H = −∑p log p; quantifies uncertainty in stimuli or responses. ↪ Information Rate
• Spike‑Timing‑Dependent Plasticity (STDP): ∆w ∝ exp(−|Δt|/τ) with sign depending on pre‑post order. ↪ Hebbian, LTP/LTD
• State Space Model: Latent variable approach describing neural dynamics trajectories. ↪ Kalman Filter
• Synaptic Scaling: Global multiplicative weight change maintaining relative differences. ↪ Homeostatic Plasticity
• Temporal Coding: Information in precise spike times or phases, not just rates. ↪ Burst Firing
• Thalamocortical Relay: Pathway gating sensory flow; burst vs tonic modes. ↪ Oscillation
• Transfer Entropy: Directed information flow metric capturing non‑linear dependencies. ↪ Effective Connectivity
• Tripartite Synapse: Functional unit including pre‑, post‑synaptic neurons and astrocyte processes. ↪ Astrocyte
• Unsupervised Learning: Self‑organization without labeled output (e.g., ICA, Hebbian). ↪ Self‑Organizing Map

V–Z

• Variance Explained (R²): Fraction of data variance captured by model; goodness‑of‑fit measure. ↪ GLM
• Variational Inference: Approximates posteriors by optimizing KL divergence; proposed for cortical computation. ↪ Free‑Energy Principle
• Voltage‑Clamp: Technique fixing membrane potential to isolate ionic currents. ↪ Patch‑Clamp
• Winner‑Take‑All (WTA): Competitive network motif where strongest unit suppresses others; decision making. ↪ Lateral Inhibition
• XOR Problem: Non‑linearly separable task illustrating need for hidden layers; milestone in connectionism. ↪ Perceptron
• Zeitgeber: External cue (light) entraining circadian rhythms; affects SCN firing. ↪ Oscillation

---

Suggested Further Reading

• Dayan & Abbott — Theoretical Neuroscience (2001)
• Gerstner et al. — Neuronal Dynamics (2014)
• Friston KJ — The Free‑Energy Principle (2010, Nat Rev Neurosci)

Need a term not listed? Ping me and we’ll expand the glossary!