Laurent Perrinet

The remarkable energy efficiency of the Human brain: One #Spike Every 6 Seconds !

In the groundbreaking paper "The Cost of Cortical Computation" published in 2003 in Current Biology, neuroscientist Peter Lennie reached a stunning conclusion about neural activity in the human brain: the average firing rate of cortical neurons is approximately 0.16 Hz—equivalent to just one spike every 6 seconds.

This finding challenges conventional assumptions about neural activity and reveals the extraordinary energy efficiency of the brain's computational strategy. Unconventional? Ask a LLM about it, and it will rather point to a baseline frequency between 0.1Hz and 10Hz. Pretty high and vague, right? But how did Lennie arrive at this remarkable figure?

The Calculation Behind the 0.16 Hz Baseline Rate

Lennie's analysis combines several critical factors:

1. Energy Constraints Analysis

Starting with the brain's known energy consumption (approximately 20% of the body's entire energy budget despite being only 2% of body weight), Lennie worked backward to determine how many action potentials this energy could reasonably support.

2. Precise Metabolic Costs

His calculations incorporated detailed metabolic requirements:

  • Each action potential consumes approximately 3.84 × 109 ATP molecules
  • The human brain uses about 5.7 × 1021 ATP molecules daily

3. Neural Architecture

The analysis factored in essential neuroanatomical data:

  • The human cerebral cortex contains roughly 1010 neurons
  • Each neuron forms approximately 104 synaptic connections

4. Metabolic Distribution

Using cerebral glucose utilization measurements from PET studies, Lennie accounted for energy allocation across different neural processes:

  • Maintaining resting membrane potentials
  • Generating action potentials
  • Powering synaptic transmission

By synthesizing these factors and dividing the available energy budget by the number of neurons and the energy cost per spike, Lennie calculated that cortical neurons can only sustain an average firing rate of approximately 0.16 Hz while remaining within the brain's metabolic constraints.

Implications for Neural Coding

This extremely low firing rate has profound implications for our understanding of neural computation. It suggests that:

  • Neural coding must be remarkably sparse — information in the brain is likely represented by the activity of relatively few neurons at any given moment
  • Energy efficiency has shaped brain evolution — metabolic constraints have driven the development of computational strategies that maximize information processing while minimizing energy use
  • Low baseline rates enable selective amplification — this sparse background activity creates a context where meaningful signals can be effectively amplified

    • The brain's solution to energy constraints reveals an elegant approach to computation: doing more with less through strategic sparsity rather than constant activity.

    This perspective on neural efficiency continues to influence our understanding of brain function and inspires energy-efficient approaches to #ArtificialNeuralNetworks and #neuromorphic computing.

    Apr 30, 2025 at 4:29pmWeb
    Show threadNicole RustMay 2, 2025
    @laurentperrinet
    I endorse! This was a terrific paper.
    Show threadDan GoodmanMay 2, 2025
    @NicoleCRust @laurentperrinet yes really spectacular result and so simply argued. And a result that neuroscience still hasn't really come to terms with in my opinion. That's a really tiny number of spikes, and the fact that it doesn't vary depending on task means that there's a very robust mechanism to ensure that. And as far as I can tell we have very little evidence what that mechanism is.
    Show threadNicole RustMay 2, 2025
    @neuralreckoning @laurentperrinet
    I agree!
    Show threadLaurent PerrinetMay 2, 2025

    @NicoleCRust @neuralreckoning

    In fact, while preparing a lecture where I've always cited "neurons spike at roughly 1Hz on average," I decided to verify this figure. Revisiting that paper from Lennie, I was genuinely surprised to find the actual rate is even lower.

    This realization coincided with my reading of an excellent recent paper on decoding information from large neural datasets: https://arxiv.org/html/2504.08201v3
    The authors apply a common preprocessing step - eliminating all neurons firing below 2Hz. This practice is standard and methodologically defensible, but it raises a critical question: by routinely filtering out these sparsely-firing neurons, might we be discarding significant information?

    If the grand average is substantially below 1Hz, then these preprocessing steps could systematically exclude entire populations of neurons whose rare but precisely-timed spikes might carry important computational significance. This points to a potential blind spot in our analytical approaches that deserves more careful consideration. (more about that in the excellent paper by Bruno Olshausen "How Close Are We to Understanding V1?" from 2005 doi:10.1162/0899766054026639 )

    Neural Encoding and Decoding at Scale

    Show threadAlexander WeberMay 2, 2025
    @laurentperrinet But is average the best way to represent neural firing if we know it follows a power law?
    Show threadLaurent PerrinetMay 2, 2025

    @weberam2 You're absolutely right! Information conveyed through neural activity is primarily encoded in modulations of firing rates at any given moment.

    However, examining long timescales for individual neurons provides valuable insights into population-level encoding. Each neuron maintains its characteristic baseline firing rate (with inhibitory neurons typically firing more frequently). This diversity is well-documented in the hippocampus, as shown in this study detailing rat hippocampal neurons during different sleep states.

    What's particularly fascinating is that finding an upper bound of just one spike every ~6 seconds for the "grand average" firing rate suggests some neurons fire extremely rarely. This challenges the standard Poisson point-process model of neural encoding and supports an intriguing alternative hypothesis: that precise spike timing itself carries significant information beyond just rate coding.

    Neuronal firing rates diverge during REM and homogenize during non-REM - Scientific Reports

    Neurons fire at highly variable intrinsic rates and recent evidence suggests that low- and high-firing rate neurons display different plasticity and dynamics. Furthermore, recent publications imply possibly differing rate-dependent effects in hippocampus versus neocortex, but those analyses were carried out separately and with potentially important differences. To more effectively synthesize these questions, we analyzed the firing rate dynamics of populations of neurons in both hippocampal CA1 and frontal cortex under one framework that avoids the pitfalls of previous analyses and accounts for regression to the mean (RTM). We observed several consistent effects across these regions. While rapid eye movement (REM) sleep was marked by decreased hippocampal firing and increased neocortical firing, in both regions firing rate distributions widened during REM due to differential changes in high- versus low-firing rate cells in parallel with increased interneuron activity. In contrast, upon non-REM (NREM) sleep, firing rate distributions narrowed while interneuron firing decreased. Interestingly, hippocampal interneuron activity closely followed the patterns observed in neocortical principal cells rather than the hippocampal principal cells, suggestive of long-range interactions. Following these undulations in variance, the net effect of sleep was a decrease in firing rates. These decreases were greater in lower-firing hippocampal neurons but also higher-firing frontal cortical neurons, suggestive of greater plasticity in these cell groups. Our results across two different regions, and with statistical corrections, indicate that the hippocampus and neocortex show a mixture of differences and similarities as they cycle between sleep states with a unifying characteristic of homogenization of firing during NREM and diversification during REM.

    Nature
    Show threadLaurent PerrinetMay 7, 2025

    another paper of interest on a similar topic:

    https://pmc.ncbi.nlm.nih.gov/articles/PMC8106317/

    • Evaluates energy consumption of human brain to around 7W (not the 20W we often repeat),
    • evaluates the ratio of energy costs of computation vs. communication to 1 over 35,
    • synaptic events needed to fire a spike are ~2000 (that is, 20% of the average 10k synapses converging on each neuron).
    Communication consumes 35 times more energy than computation in the human cortex, but both costs are needed to predict synapse number

    Engineers describe the human brain as a low-energy form of computation. However, from the simplest physical viewpoint, a neuron’s computation cost is remarkably larger than the best possible bits per joule—off by a factor of 108. Here we explicate, ...

    PubMed Central (PMC)