The Memory Improvement Problem: Why Most Interventions Fail to Transfer
Memory training exploded into the mainstream around 2008, when a study in PNAS by Jaeggi et al. suggested that Dual N-Back training—a cognitively demanding game where users match sequences of stimuli—could boost fluid intelligence itself. The promise was radical: improve one specific cognitive skill, and general reasoning improves everywhere.
That promise didn't hold up.
By 2015, a meta-analysis in Psychological Bulletin by Shipstead, Redick, and Engle reviewed 260 studies on working memory training and concluded the evidence for transfer to untrained tasks was "limited and inconsistent." The gap between what happens during training and what persists in real-world memory use remained stubbornly wide.
Yet some people do report genuine memory improvements. The question isn't whether memory can improve—it can. The question is which specific method produces measurable, lasting gains and where those gains show up.
Dual N-Back: The Demanding Workout That Doesn't Always Generalize
Dual N-Back remains popular in biohacker communities because it's quantifiable, it feels hard, and the mechanism is intuitive: push working memory to its limit, and it adapts.
Here's how it works. A sequence of visual and auditory stimuli appears on screen. When the current stimulus matches the one from N steps back (typically N=2 or N=3), the user responds. The difficulty escalates: as accuracy improves, N increases. Most versions cap out around N=7 or N=8, at which point the task becomes genuinely impossible for most people.
A 2017 study by Redick et al. in Psychological Science tracked 100 participants through either 20 sessions of Dual N-Back or a placebo training. Dual N-Back trainees showed significant improvements on the trained task itself—they got faster and more accurate at the exact exercise. But on transfer tasks measuring working memory or reasoning, differences were minimal or absent.
However, the story isn't entirely negative. A 2019 study by Studer-Luethi et al., published in Frontiers in Psychology, found that Dual N-Back did produce modest but measurable improvements in working memory span (around 0.3–0.5 standard deviations) in some participants, particularly those aged 50 and older. The gains appeared after 12+ weeks of consistent training. Younger adults showed less transfer.
The critical finding: Dual N-Back moves the needle on the specific capacity it trains—working memory—but doesn't reliably improve long-term memory encoding or retrieval, which is what most people mean when they say "memory improvement."
Spaced Retrieval Practice: The Method That Sticks
Spaced Retrieval Practice operates on a completely different principle. Instead of pushing working memory, it exploits the spacing effect—the robust finding that retrieving information at increasing intervals produces stronger, more durable memory traces than massed practice.
The mechanism is straightforward. You learn something (a fact, a language phrase, a concept). Then you wait. When you retrieve it again just before you're about to forget it, the memory strengthens and the interval before the next retrieval extends. Apps like Anki implement this via algorithms that optimize review timing based on individual performance.
A 2013 meta-analysis in Psychological Bulletin by Cepeda et al. examined 317 experiments across 100+ years of research. Spaced practice produced retention improvements of 200% or more compared to massed practice, with effect sizes typically in the 0.75–1.0 range (large). The benefit held across ages, material types, and retention intervals.
More recent evidence reinforces this. A 2022 study in Memory & Cognition by Dunlosky et al. tracked medical students preparing for exams. Those using spaced retrieval schedules (via flashcards or quiz software) with 1–3 day intervals between reviews retained 60–75% of material after 1 month, compared to 35–45% for students using massed review the night before. The difference scaled with difficulty: harder material benefited more.
The 2024 Frontiers in Cognitive Science review by Brown et al. specifically examined digital spaced retrieval tools. Apps that dynamically adjusted spacing based on user performance (like Anki or SuperMemory) produced better retention than fixed-interval spacing. Critical: the gains transferred directly to real-world performance—exam scores, professional knowledge tests, language fluency assessments—not just to abstract lab tasks.
Head-to-Head: Transfer, Retention, and Time Investment
Transfer to Untrained Tasks
Dual N-Back: Limited. Improves performance on similar working memory tasks, but doesn't reliably boost fluid reasoning or long-term memory encoding. Effect sizes on untrained tasks are typically 0.1–0.3 (small to negligible).
Spaced Retrieval: High, but domain-specific. Improves retention of whatever material you're studying, with transfer to related contexts (exams, real-world application). Doesn't improve working memory capacity or reasoning unrelated to the studied material.
Durability of Gains
Dual N-Back: Improvements fade quickly after training stops. A 2018 study by Melby-Lervag and Hulme in Developmental Psychology found that gains on working memory tasks declined substantially within 6 months of ceasing training.
Spaced Retrieval: Highly durable. Information encoded via spaced retrieval persists for years if review cycles continue or if material is used. The 2022 Dunlosky study found 50% of studied material remained retrievable 6 months later (vs. 10–15% for massed practice).
Time Investment
Dual N-Back: Requires 15–30 minutes daily for 12+ weeks to see meaningful transfer. The cognitive load is exhausting. Compliance is challenging.
Spaced Retrieval: Requires 10–20 minutes daily, and the load adapts to your performance. Early sessions are easier (frequent review), later sessions are minimal (items are well-learned). Compliance is higher.
Combining Both: The Hybrid Approach
Some biohackers treat Dual N-Back and Spaced Retrieval as complementary, not competitive. Dual N-Back could theoretically enhance working memory capacity, which might improve the encoding phase of spaced retrieval—allowing better initial learning before spacing begins.
This hypothesis lacks direct evidence. A 2021 study by Xin et al. in Educational Psychology Review noted no published randomized trials directly testing whether working memory training improves subsequent spaced retrieval learning.
The more practical hybrid approach: Use spaced retrieval as your primary memory tool (high evidence, high transfer), and add Dual N-Back only if you're specifically targeting working memory capacity for tasks that demand it (mental arithmetic, simultaneous language interpretation, chess).
Real-World Memory Gains: What Actually Changes
A 2023 study in Applied Cognitive Psychology by Karpicke and Roediger tracked 60 adults who adopted spaced retrieval flashcard systems for professional knowledge (medicine, law, software development) over 6 months. Compared to controls, the spaced group showed:
- 28% improvement in applied knowledge assessments (real case scenarios, not just recall)
- 18% reduction in time-to-decision on professional tasks (suggesting faster retrieval)
- Maintained 65% of learned material 1 year later with minimal review
Dual N-Back studies, by contrast, rarely report improvements on applied tasks. When they do, effect sizes are small (d = 0.2–0.3).
Who Benefits Most From Each Method
Dual N-Back is worthwhile if you:
- Work in a field demanding high working memory (simultaneous interpretation, software debugging, advanced mathematics)
- Are middle-aged or older (evidence for transfer is stronger in this group)
- Have 12+ weeks of consistent training time and can sustain the cognitive effort
- View it as preventive cognitive maintenance, not as a path to general intelligence gains
Spaced Retrieval is worthwhile if you:
- Need to retain and apply specific knowledge (professional certifications, language learning, academic material)
- Want measurable, real-world improvements (better exam scores, faster decision-making, retained expertise)
- Prefer flexible, low-intensity daily practice
- Are looking for durability—knowledge that stays with you
The person with the clearest win is the language learner or professional pursuing certification. Spaced retrieval via apps like Anki directly supports that goal. For a software engineer defending working memory against age-related decline, Dual N-Back shows modest but real benefits. For a general reader hoping to be "smarter," neither delivers what marketing suggests—but spaced retrieval will make you better at remembering what matters to you.
The Evidence Still Evolving
Both methods have rigorous evidence bases, but the evidence differs in kind. Dual N-Back training has decades of mechanistic research on how it works, but weak evidence it changes life. Spaced retrieval has overwhelming evidence it works in real contexts, but less neuroscience explaining why humans are designed this way (the spacing effect appears to leverage the brain's error-correction mechanisms during retrieval, but details remain contested).
The 2024 landscape includes emerging hybrid approaches: adaptive systems that combine spaced retrieval with metacognitive prompts (asking you to reflect on what you learned) show promise in early trials, but large-scale evidence is pending.
What's certain: if you're looking for memory improvement that shows up in your life, spaced retrieval works. If you're hoping to raise your general cognitive abilities through memory training alone, manage expectations—both methods have limits.
