The Spike: An Epic Journey Through the Brain in 2.1 Seconds
Overview
In The Spike, computational neuroscientist Mark Humphries follows a single moment of perception — catching a falling cup — through the brain in 2.1 seconds, using the electrical spike as the narrative thread. The book explains how neurons generate, transmit, and decode spikes, then confronts the puzzles that simple input-output models cannot explain: the prevalence of “dark” silent neurons, the impossibility of sustained high firing rates, and the ease with which single stimuli cross threshold. Humphries resolves these by connecting them to spontaneous activity, population coding, and predictive processing.
Key Concepts
The Spike as Neural Currency
- Action potential mechanics: A neuron fires when its membrane potential crosses a threshold (~-55 mV), producing an all-or-nothing electrical pulse that propagates down the axon. The spike itself carries no graded information — all the meaning lies in when it fires and in combination with whom.
- Synaptic transmission: At the synapse, the spike triggers neurotransmitter release. Excitatory neurotransmitters (glutamate) push the receiving neuron toward threshold; inhibitory ones (GABA) pull it away. The receiving neuron integrates thousands of such inputs before deciding whether to spike.
- Rate vs. temporal coding: The classical view holds that information is carried by firing rate (spikes per second). Humphries shows that timing matters too — precise spike timing relative to other neurons and to oscillatory rhythms carries additional information.
The Dark Neuron Problem
- Silent majority: In any given moment, the vast majority of cortical neurons are not firing. These “dark neurons” appear inactive when we record from them during experiments, yet they constitute most of the brain.
- Spontaneous activity: Dark neurons are not truly silent — they maintain a low-level spontaneous activity that keeps their membrane potential close to threshold, ready to fire rapidly when needed. This resolves the paradox of how the brain responds so quickly to new stimuli.
- Sparse coding: Only a small, shifting subset of neurons fires at any time, which is metabolically efficient and increases the brain’s representational capacity (more possible patterns from the same neurons).
Excitation-Inhibition Balance
- The E/I dance: Cortical circuits maintain a tight balance between excitatory and inhibitory activity. Inhibitory interneurons act as a brake, preventing runaway excitation (which would produce seizures) while sculpting the precise timing of population activity.
- Winner-take-all dynamics: Inhibition enables competition — when one group of neurons fires, it suppresses alternatives, sharpening the neural representation and enabling decision-making.
Predictive Processing and Anticipation
- Expectation as background activity: Humphries links the spontaneous, sub-threshold activity of dark neurons to the brain’s predictive machinery. The brain is not waiting passively for input — it is constantly generating expectations about what will happen next.
- Prediction error signals: When sensory input matches the prediction, little additional spiking is needed. When it deviates, a burst of prediction-error spikes propagates up the hierarchy to update the internal model. This explains why unexpected stimuli produce stronger neural responses.
- Speed of processing: Because predictions pre-activate the relevant neural populations, the brain can respond to stimuli in fractions of a second — far faster than if it had to build representations from scratch.
Personal Reflection
[To be added]
Related Books
- Being You - Seth and Humphreys both build on prediction and neural coding; complementary perspectives on consciousness
- Active Inference - Parr et al.’s formal framework gives mathematical depth to the predictive-brain ideas Humphreys narrates
- Neuropedia - Coltart’s quick-reference entries contextualise the neurons and circuits Humphreys describes
Parent: Books