The Science of Spaced Repetition and Memory

Quick Summary: Spaced Repetition (Spaced Repetition System, SRS) is a learning method backed by over a century of cognitive science and neuroscience research. You review material at progressively wider intervals, each time just before you’d forget it, and that activates Long-Term Potentiation (LTP) in your synapses. Fragile short-term memories become durable knowledge, with over 90% retention in far less study time than traditional massed study.

---

The Ebbinghaus Forgetting Curve: The Mathematics of Memory Decay

In 1885, German psychologist Hermann Ebbinghaus ran the first quantitative investigation of human memory retention (Memory: A Contribution to Experimental Psychology). He found that after you learn something new, memory decays exponentially:

$$\text{Retention }R=e^{-\frac{t}{S}}$$

Where:

• R: probability of recalling the information.
• t: time since the last study session.
• S: memory stability (how well the neural trace resists decay).

Average Memory Decay Without Spaced Practice:
100% ── Initial Learning Event
 50% ── Forgotten after just 1 HOUR
 70% ── Forgotten after 24 HOURS
 80% ── Forgotten after 1 WEEK

That’s why last-minute cramming can get you through tomorrow’s exam and leave you with almost nothing a few weeks later. For more, see why traditional study techniques fail.

---

The Neurobiological Basis: Synaptic Plasticity, LTP, and Desirable Difficulties

To see why spacing works so well, look at learning at the cellular level:

1. Long-Term Potentiation (LTP) and Synaptic Plasticity

When two neurons fire together again and again during retrieval, the connection between them physically strengthens as AMPA and NMDA glutamate receptors are recruited. This is Long-Term Potentiation (LTP).

Research shows that stimulation spread out over time recruits more synaptic proteins and triggers more sustained gene expression than continuous, massed stimulation.

2. The Desirable Difficulties Framework (Robert Bjork)

UCLA’s Robert Bjork showed that durable learning needs Desirable Difficulties:

• Easy now, forgotten fast: rereading or looking at highlighted text takes little effort and gives you an illusion of competence.
• Hard now, lasting memory: making your brain retrieve information (Active Recall) right before it slips away creates the effort that encodes lasting neural pathways.

---

The Algorithmic Evolution: From Leitner (1972) to FSRS

Spaced repetition technology has evolved over five decades to predict, mathematically, when each review should happen:

Era / Algorithm	Year	Core Scheduling Mechanism	Major Constraints
Leitner System	1972	Physical card boxes with fixed step intervals	Inflexible; cannot adapt to varying card difficulty
SuperMemo SM-2	1987	Ease Factor multipliers based on performance	Susceptible to the punishing Ease Hell trap
Half-Life Regression (HLR)	2016	Continuous memory half-life modeling via ML	Built for closed language curricula (Duolingo)
FSRS (Free Spaced Repetition)	2023	DSR Framework (Difficulty, Stability, Retrievability)	Requires manual tuning of multiple parameters
What Affex runs	2026	SM-2 or Leitner for flashcards; simplified HLR for course exercises	No FSRS yet; works out of the box

---

Synaptic Consolidation and Memory Flow Diagram

flowchart TD
    A[Initial Learning: Fragile Neural Trace Formed] --> B[Exponential Memory Decay / Ebbinghaus Curve]

    B --> C{SRS Review Trigger Point}

    C -- Review at Optimal Point of Near-Forgetting --> D[Effortful Active Retrieval Practice]
    D --> E[LTP Synaptic Potentiation Triggered]
    E --> F[Stability S Increases: Curve Flattens Significantly]
    F --> G[Permanent Consolidation into the Neocortex]

    C -- No Spaced Review --> H[Synaptic Pruning: Information Discarded as Neural Noise]

With each successful review in Affex, the forgetting curve flattens and the interval until the next review grows exponentially (1 day, 4 days, 14 days, 60 days, 6 months).

---

Scientific Evidence: What Meta-Analyses Prove

The advantage of distributed practice and active testing is among the most replicated findings in psychology:

• Cepeda et al. (2006, Psychological Bulletin): reviewed 254 empirical studies covering over 14,000 participants and concluded that the spacing effect gives large, statistically significant memory gains across age groups and subjects.
• Dunlosky et al. (2013, Association for Psychological Science): ranked active retrieval (Practice Testing) and spaced distribution (Distributed Practice) as the only two techniques with “High Utility” among all the mainstream methods evaluated.
• Karpicke & Blunt (2011, Science): students who used active retrieval retained 50% more conceptual information on tests than students who used concept mapping or repeated rereading.

---

How Affex Applies Cognitive Science in Practice

Many study tools implement only part of the picture. Affex puts it together in one workflow:

1. Mandatory Active Recall: cards make you produce the answer mentally before you flip.
2. Dual scheduling: a simplified Half-Life Regression for structured courses and progressive lessons, and SM-2 (or Leitner) for dedicated flashcards.
3. Zero configuration: Anki asks you to tune parameters by hand, and Quizlet keeps SRS behind a paywall. Affex schedules with SM-2 out of the box, for free.

To bring this into your routine, read the guide to 11 practical ways to apply spaced repetition or the market comparison in Affex vs. Alternatives.

---

Conclusion

Studying by rereading and massed repetition is like “renting” memory that disappears right after the test. With evidence-based spaced repetition, you build connections that last for years.

---

Experience Affex’s spaced repetition algorithm for free →