There is a common misconception about mental exercises: the idea that any cognitive activity improves cognition generally. Research points to a more specific reality — each type of exercise strengthens the system it trains, with limited transfer to other systems. This does not diminish the value of practice; it changes where to direct the effort.
The brain has distinct cognitive systems — working memory, selective attention, language, executive function, spatial perception — that respond to different stimuli and do not automatically strengthen one another. An adult who does crosswords daily may improve vocabulary and semantic retrieval without meaningful gains in visual memory. A memory-game player may develop spatial strategies without becoming better at verbal reasoning.
What each cognitive system does — and which game recruits it
Visuospatial working memory: Baddeley's "sketchpad" component, which sustains tasks like reading a map, assembling an object or tracking a visual sequence. Super Zoo recruits this directly: on Hard mode with 23 pairs, the player must maintain a mental map of up to 46 card positions without external support. The chunking strategy players spontaneously develop — dividing the grid into regions — mirrors the working-memory chunking described in cognitive literature.
Selective attention and visual search: Word Search trains conjunctive visual search — finding a target that combines multiple attributes (specific letter + sequence + direction) among distractors. This is the same process used to scan a document for a specific term or filter relevant information from a complex table.
Semantic memory through associative networks: Crosswords activate convergence reasoning — the brain simultaneously traverses semantic networks (capitals + 8 letters + starts with B) until the intersection node is found. This activates more pathways than passive reading and keeps semantic networks more accessible.
Active retrieval and long-term memory consolidation: Educational Quiz operates through the testing effect documented by Roediger and Karpicke: being tested on content consolidates its memory more durably than studying it one more time. Even making a mistake and seeing the correct answer immediately produces a stronger memory trace than simply reading the right answer.
Verbal working memory and numerical reasoning: Math Adventure recruits the phonological loop — the system that holds auditory/verbal information active long enough to be manipulated. Mental calculation requires maintaining intermediate results while executing the next operation.
Executive function and inhibitory control: Enchanted Maze recruits the central executive — maintaining the goal (exit the maze) while suppressing automatic responses (go straight) and replanning in real time.
Why "training the brain" generically does not work
Literature reviews — including a statement signed by over 70 neuroscientists in 2014 — document that computerised cognitive training rarely produces benefits beyond the specific task type being trained. The practical implication: vary across genuinely different types of exercise, because each modality recruits systems the others do not.
Conclusion
The recommendation to vary mental exercises is not vague common sense — it is grounded in the fact that visuospatial memory, selective attention, semantic memory, active retrieval and executive function are distinct systems that do not cross-train. Super Zoo trains what Word Search does not; Crosswords activate what Educational Quiz does not in the same way. A routine that alternates between these games covers a cognitive spectrum that none of them covers alone.