Just a teeny, tiny taste of what I found:
Here’s a deep dive into how predictive coding and timing mechanisms shape human perception of time—the sense of “now,” the misalignment of events, and temporal illusions like the flash‑lag, kappa, and tau effects.
1. Real‑time Extrapolation & the Flash‑Lag Illusion
-
Flash‑lag illusion: A moving object appears ahead of a simultaneously flashed object because the brain extrapolates motion forward, compensating for neural delays (~100 ms).
- Motion‑based predictive model accurately reproduces human psychophysical data by combining current sensory data with motion priors and explicit delay compensation jov.arvojournals.org+14journals.plos.org+14en.wikipedia.org+14.
-
Delay‑aware predictive coding model implements both forward (extrapolation) and backward (alignment) processes to realign predictions across a cortical hierarchy en.wikipedia.org+2eneuro.org+2nature.com+2.
➤ Result: The brain’s perceptual “now” is a prediction-corrected present, not simply the raw sensory input.
2. Time‑binding Illusions: Tau & Kappa Effects
- Tau effect: When the interval between stimuli varies, the perceived spatial distance is warped—time influences space perception en.wikipedia.org+1en.wikipedia.org+1.
-
Kappa effect: Longer spatial separation makes intervals feel longer; shorter ones feel shorter—echoing spatiotemporal Bayesian inference with velocity priors .
➤ These illusions reveal that perception of time and space is interwoven, jointly inferred under predictive coding frameworks.
3. Temporal Asynchrony in Visual Features
-
Perceptual asynchrony: Color changes are perceived ~70–80 ms before motion changes in the same object researchgate.net+15en.wikipedia.org+15pmc.ncbi.nlm.nih.gov+15.
➤ This indicates the brain has feature-specific processing latencies, yet we experience a unified moment through temporal integration and prediction.
4. Temporal Scales in Hierarchical Predictive Coding
-
Hierarchical time‑scales: Cortical levels track different temporal extents; lower levels respond to fast-changing inputs while higher levels integrate over seconds en.wikipedia.org+1en.wikipedia.org+1.
➤ The brain maintains multiple “nows”, enabling both immediate sensory processing and context-sensitive perception.
5. Neural Oscillations & Temporal Prediction
- Oscillations (e.g., alpha, gamma) structure time into meaningful chunks, enabling prediction and error signaling at specific phases en.wikipedia.org.
-
Theta–gamma coupling in the hippocampus compresses sequences so STDP can link temporally distant events into coherent episodes .
➤ The brain uses rhythms as clocks, organizing prediction and sensory integration within oscillatory cycles.
-
Theta–gamma coupling in the hippocampus compresses sequences so STDP can link temporally distant events into coherent episodes .
6. EEG/FMRI Evidence of Predictive Temporal Processing
- Prediction error ERPs (e.g., Ne/ERN) reflect fast mismatch detection (~100 ms), even for timing violations sciencedirect.com+15en.wikipedia.org+15nature.com+15arxiv.org+4pmc.ncbi.nlm.nih.gov+4journals.physiology.org+4.
-
Early vs late ERP components correspond to feature‑level mismatches vs identity/context mismatches in visual prediction tasks pmc.ncbi.nlm.nih.gov+1sciencedirect.com+1.
➤ Temporal prediction is layered: early feedforward errors, then recurrent identity/contextual updating.
Bringing It Together: Time in Human Experience
- Perception of “now” is a reconstruction, synthesized from momentary inputs and hierarchical predictions.
- Temporal illusions like tau/kappa occur because the brain merges space and time under Bayesian expectations.
- Feature asynchrony reflects modular timing but unified experience emerges via synchronization.
- Oscillatory rhythms gate when errors are computed and changes are integrated.
- ERP evidence shows timing prediction errors manifest rapidly and hierarchically in the brain.
Summary
Phenomenon | Mechanism | Perceptual Outcome |
---|---|---|
Flash-lag | Motion extrapolation + delay alignment | Moving objects appear ahead |
Tau/Kappa | Bayesian expectation of uniform motion | Spatiotemporal illusions |
Feature asynchrony | Modular processing latencies + integration | Integrated “now” experience |
Hierarchical timescales | Layered predictive hierarchy | Fast perception + slow contextual updating |
Oscillations | Temporal gating | Rhythmic timing for prediction & error |
ERPs | Early vs late prediction errors | Layered temporal mismatch detection |