Cognitive Warm-Up

Before a student begins reading tasks, the platform runs a short warm-up exercise that has nothing to do with reading. It is there to settle attention, prime working memory, and reduce the cognitive cost of starting. A student who has just switched contexts, walked in from recess, or had a moment of distraction can re-focus through the warm-up rather than burning that re-focus energy on the first measurement task.

What it is

The warm-up is a sequence of visual exercises. Students see shapes, colors, sequences, and spatial patterns. They respond to what they see. There are no letters, no words, no numbers, and nothing that draws on language skills. This is not accidental: if the warm-up required reading, it would tire the skills the platform is about to measure.

The exercise types address four cognitive areas: attention and inhibition, pattern recognition, working memory, and visual-spatial reasoning. Which types appear and how difficult they are depends on the student’s grade. A first-grader sees simpler patterns; a fourth-grader encounters longer sequences and more demanding spatial tasks.

The warm-up takes between three and five minutes under normal conditions. The platform enforces a hard cap of five minutes. If a student is still mid-exercise when the cap is reached, the session advances automatically without penalty. A student who needs to stop early or skip entirely does not face any consequence. There is no “warm-up streak,” no points for completing it, and no negative signal from skipping.

Why it is completely separate from measurement

This point is load-bearing: the warm-up has no connection to reading assessment. Its results are never factored into a reading-ability estimate. They do not contribute to any skill status, domain status, or student profile. They are not included in the evidence that drives a support recommendation. The warm-up’s only output is a readiness signal stored in its own isolated tables, visible only to platform engineers for diagnostic purposes.

The code enforces this separation structurally. The warm-up module has no database link into the measurement or decision tables. A warm-up completion event cannot write into the rows that hold reading results. This is tested in the build as a structural isolation check.

Similarly, the warm-up is not driven by a language model. It is rule-based and deterministic: given the same grade level and the same starting seed, the same sequence of exercises is always produced. This means any warm-up session can be replicated exactly for debugging or review, without relying on inference.

Resumability

A student who is interrupted mid-warm-up can resume where they left off. If a browser tab is closed, the device loses connection, or the student switches apps, the platform restores the session to the exact exercise the student was on.

This restoration is reliable because the warm-up state is persisted on the server. The client receives a state token on session creation that it uses to request a resume. The resume path regenerates the exercise deterministically from the stored position: no data needs to be re-fetched, and no exercise is repeated or skipped because of the interruption.

The five-minute cap counts from the original session start time, not from the resume. A student who spent three minutes before an interruption has two minutes remaining when they resume.

What teachers see

Teachers do not see warm-up performance. There is no warm-up score, no readiness rating, and no warm-up history in a student’s profile. The warm-up is deliberately invisible to the instructional layer.

If a student consistently skips the warm-up, there is no flag for this in the teacher’s view. The warm-up is an opt-in preparation step, not a required activity that teachers monitor.

The exercise types (Wave 1)

Ten exercise types are active in Wave 1. Two additional types (a visual-matching task and a visual-tracking task) are built and tested but remain inactive pending partner confirmation of the per-grade configurations. When those are activated, no code change is required; it is a configuration update.

The ten active types span pattern completion, color and shape sequences, memory-span tasks, and spatial-orientation exercises. Each type has its own difficulty progression across grades. A single exercise within a type is generated from a fixed template, a grade-appropriate configuration, and a deterministic seed, so the output can always be reproduced.

Assets in the warm-up are referenced by identifier, not embedded. The actual SVG images are a separate deliverable; in Wave 1, placeholder assets serve as stand-ins for exercises that require specific imagery. Production artwork and audio assets are part of the front-end and content delivery work.

Where to go next

• Accessibility explains how the warm-up handles device diversity and connection reliability.
• Diagnostic & Practice describes what follows the warm-up: the adaptive reading check itself.