The science behind how Synapse teaches.
Synapse isn't a productivity gimmick. Every review, quiz, and flashcard is wired to a principle from 140 years of memory research. Here are the six that matter — what the evidence says, and exactly where each one shows up in the product.
Free plan included · No credit card required
Built on how memory actually works
Each principle below has decades of experimental evidence behind it — and a concrete feature in Synapse.
Spaced repetition
Time your reviews so memory never fully fades.
In 1885, Hermann Ebbinghaus charted the first forgetting curve: memory for new material decays fastest right after learning, then levels off. The curve looks inevitable — until you review. Each well-timed review doesn't just restore the memory; it flattens the next curve.
The decisive evidence came from Cepeda and colleagues, whose meta-analysis covered 839 assessments across 317 experiments: spreading the same total study time over several sessions reliably beats cramming it into one, especially when the goal is remembering weeks or months later. Their follow-up found the optimal gap is roughly 10–20% of how long you want to remember — review too soon and you waste time, too late and it's already gone.
This is why "I'll re-read it the night before the interview" fails, and why five minutes of review today outperforms an hour next month.
Evidence
Memory: A Contribution to Experimental Psychology — the original forgetting-curve experiments.
Meta-analysis of 317 experiments: distributed practice beats massed practice for long-term retention.
The optimal gap between study and review is ~10–20% of how long you want to retain.
How Synapse applies it
- Every topic you complete enters a review schedule of Leitner boxes at 1, 3, 7, 14, 30, and 90 days.
- Recall it and the topic moves to a longer interval; miss it and it drops back — the gap grows only as fast as your memory proves it can.
Retrieval practice
Pulling knowledge out beats pushing it in again.
Rereading feels productive and testing feels stressful — intuition has it exactly backwards. Roediger and Karpicke had students study prose passages and then either reread them or take recall tests. A week later, the tested group remembered dramatically more, even though the rereaders had predicted the opposite.
This testing effect is one of the most replicated findings in the field. The act of retrieving strengthens the retrieval route itself and slows forgetting, while restudying only adds fluency. In Dunlosky and colleagues' landmark review of ten learning techniques, practice testing earned their highest utility rating — alongside distributed practice.
Retrieval also does the calibration job: every failed recall shows you precisely where the gaps are.
Evidence
The landmark testing-effect study: recall tests beat restudy one week later.
Review of a century of evidence on why testing memory improves it.
Rates practice testing (with distributed practice) as the only high-utility techniques.
How Synapse applies it
- Every topic ends in a quiz, and topics hand you flashcards that make you answer before they explain.
- Coding challenges ask you to produce working code, not read it — retrieval for the skills, not just the facts.
Calibrated self-rating
Know what you actually know — not what feels familiar.
Learners are notoriously poor judges of their own understanding. Rereading a solution feels like learning because it's fluent — but fluency isn't knowledge. Dunlosky and Rawson documented these illusions of competence: students studying flashcards consistently predicted they had learned far more than their later tests showed.
The cost is real. In their classroom study, overconfident students quit studying too early and retained less — inaccurate self-evaluations directly undermined learning. Bjork, Dunlosky, and Kornell's review frames accurate self-monitoring as the engine of self-regulated learning: what you choose to study next only works if your picture of what you don't know is honest.
The fix is to make judgment meet reality: rate your confidence before checking the answer, then let actual performance correct the rating.
Evidence
Illusions of competence: learners overestimate what studying has fixed.
Overconfidence produces underachievement — inaccurate self-evaluations hurt retention.
Self-regulated learning: beliefs, techniques, and illusions — monitoring drives everything.
How Synapse applies it
- When a topic review is due, you rate how well you recalled it before the schedule commits.
- An honest "shaky" drops the topic back to an earlier box — overconfidence can't skip the schedule.
Interleaving
Mix it up, so you learn to tell things apart.
Blocking — drilling one kind of problem until it's exhausted before starting the next — feels efficient. Kornell and Bjork showed the opposite for learning to tell categories apart: people who mixed examples from different painters learned to classify artists' styles far better than those who studied one painter at a time. Most participants were convinced blocking had worked better.
Rohrer and Taylor found the same in mathematics: students who practiced shuffled problem sets outperformed those who worked through textbook blocks — even though blocked practice looked better during practice itself. Rohrer, Dedrick, and Stershic later replicated the effect in real middle-school classrooms.
The mechanism: mixing forces you to first ask "which kind of problem is this?" — exactly the skill an interview demands, where problems never arrive labeled.
Evidence
Interleaved examples beat blocked ones for category learning — spacing is the friend of induction.
Shuffling mathematics problems improves learning over blocked practice.
Interleaved practice improves mathematics learning in real classrooms.
How Synapse applies it
- Review sessions pull topics from across your whole course instead of drilling one at a time.
- The practice lab shuffles challenge types, so pattern-matching can't substitute for thinking.
Elaboration / Feynman
If you can't explain it simply, you haven't learned it.
Explaining something in your own words forces the gaps into the open. In the research literature this appears as elaborative interrogation — asking "why is this true?" — and self-explanation — connecting new material to what you already know. Dunlosky and colleagues rated both as cheap, reliable techniques, strongest when you have some background to hook them onto.
The Feynman technique — explain the idea in plain language as if teaching it, then study wherever you stumble — is the popular packaging of the same mechanism. It hasn't been tested under that name, but it's self-explanation with the honesty turned up.
Writing beats highlighting because it's generative: you have to choose the words, and every choice exposes what you can't yet say.
Evidence
Elaborative interrogation and self-explanation: moderate-utility techniques with strong lab support.
Strengthening the Student Toolbox — the accessible practitioner summary of what works.
How Synapse applies it
- The recall journal asks you to explain each topic in your own words before you mark it reviewed.
- A per-topic chat lets you interrogate exactly the part you can't yet articulate — and the course builds to interviews where explaining is the test.
Consolidation
Learning doesn't stop when you close the book.
New memories are fragile at first and stabilize over hours — a process called consolidation, much of it during deep sleep, when the hippocampus replays the day's traces to the cortex for long-term storage. Rasch and Born's review makes the case that sleep plays an active role in this, not just a protective one.
Walker and Stickgold documented sleep-dependent improvement: on motor and perceptual skills, people get measurably better after a night of sleep without any further practice. Stickgold's review extends the picture to facts and vocabulary — sleep after studying helps you keep them.
The practical upshot: the interval between study sessions isn't dead time — it's when the learning gets finished. A schedule that respects nights beats one that doesn't.
Evidence
About sleep's role in memory — the definitive review of active systems consolidation.
Sleep-dependent learning and memory consolidation: overnight gains without further practice.
Sleep-dependent memory consolidation and reconsolidation.
How Synapse applies it
- Review gaps deliberately span nights — what you study today is due for review tomorrow morning.
- The daily plan caps new topics, so sleep always has a manageable amount to consolidate.
Read the research yourself
Every claim on this page traces back to the primary literature. For one accessible overview of which techniques actually work, start with Dunlosky et al. (2013) and the book Make It Stick.
Study the way memory works.
Synapse sequences your course, quizzes you at the right moments, resurfaces topics before they fade, and asks you to explain what you know — six principles, one habit.
Free plan included · No credit card required