How to practice Change Detection

What each level changes

Easy uses 3, 4 and 5 colored circles. A Change trial alters one color while shape and position remain fixed. The longer study display keeps the first level focused on learning the memory–delay–probe rhythm.

Standard uses 4, 5 and 6 colored shapes. A single object can change color or shape. You therefore have to remember object identity rather than only one feature family.

Challenge uses 5, 6 and 8 color–shape conjunctions, a shorter study display and a longer retention interval. A Change trial still alters exactly one feature of one object; difficulty comes from higher load and less viewing time, not hidden rules.

Same and Change are balanced

Every difficulty is 50% Same and 50% Change. This makes “always press Different” or “always press Same” an obviously poor strategy and lets the result report the two conditions separately. A Same miss means a false alarm; a Change miss means the changed object was not detected.

Use positions as anchors

The occupied positions do not move between memory and probe. That is intentional. A useful encoding strategy is to treat the 4×4 board as a small map: top-left group, center group, lower-right group, and so on. Instead of trying to rehearse eight unrelated objects as one list, bind each object to a location.

For Easy, color may be enough: “red upper left, blue center, green lower right.” Standard and Challenge benefit from paired labels such as “blue triangle at top” or “purple diamond lower left.” The goal is not to invent a perfect verbal code; it is to make each object distinct enough that the later probe can be compared with a remembered anchor.

Compare the probe deliberately

When the probe appears, avoid scanning randomly. Start with the position that feels most uncertain, then move through a consistent row or quadrant order. If one object clearly differs, respond Different. If nothing differs after one systematic pass, Same becomes a reasoned conclusion rather than a guess.

On large Challenge arrays, repeated re-checking can waste time. A stable scan order helps you know which positions have already been compared. The local set-size timing is useful here: if accuracy stays stable while response time rises with larger arrays, you may simply be taking a more careful comparison pass.

Read errors by condition

Overall accuracy alone hides two different mistakes. Low Change accuracy means real changes are being missed. Low Same accuracy means unchanged probes are being called Different. The first pattern may suggest weak encoding or incomplete probe comparison; the second can reflect uncertainty that turns into false alarms.

The game therefore keeps Same and Change accuracy separate. It also keeps median correct response time by set size. These are local session measurements, not normative memory-capacity scores.

Set size is a load control

Set size is the number of objects you must maintain across the delay. Compare performance within one difficulty first. Easy 3-versus-5 is a cleaner comparison than Easy 3-versus-Challenge 8 because Challenge also changes study time, delay and feature complexity.

If the larger set size reduces accuracy sharply, slow down the initial encoding pass. If accuracy stays high but response time rises, the memory may be adequate while the probe comparison takes longer. If both accuracy and speed collapse, step down a level before trying to accelerate.

Study Boost is assistance

Study Boost extends the memory-array exposure time. It can help you learn a stable encoding strategy, but it changes the task. Once Study Boost is used, the run is Assisted. Assisted runs stay visible in local history but cannot replace a Clean best.

A practical routine is to use one Assisted block to develop a location-based encoding method, then turn Study Boost off and repeat the same deterministic seed. That gives you the same concrete arrays under Clean timing and makes the comparison more informative than switching to unrelated material.

Repeat the same block when testing a strategy

Practice URLs carry a deterministic seed. Keeping the seed reproduces the same source and concrete memory/probe arrays. Fresh block creates a new seed. Exact replay is useful when you want to compare an encoding strategy without wondering whether the next block happened to contain easier arrays.

Keyboard and touch

Same and Different are real buttons for pointer and touch input. During the probe, S answers Same and D answers Different. Inputs are locked during the study, delay and inter-trial phases so an early double-tap cannot accidentally answer the next probe before it appears.

A useful progression

Start with Easy until the three phases feel automatic. Move to Standard and focus on remembering color and shape together. Move to Challenge only when you can maintain a systematic encoding and comparison routine without turning the final response into a guess.

When practice becomes sloppy, reducing difficulty is more useful than chasing a faster reaction time. Correct change detection depends on the stored array first; speed only has meaning after the memory representation is reliable.

Build an encoding pass before chasing speed

A reliable block starts during the study display. On smaller arrays, make one deliberate pass across occupied positions rather than darting back and forth. On larger arrays, split the board into two or four regions and encode each region once. The point is to create a stable memory trace before the screen disappears, not to squeeze in as many eye movements as possible.

If you know a particular position is weak, do not spend the entire study interval staring at it. Finish the whole array first, then use any remaining time for one targeted re-check. This avoids the common tradeoff where one object becomes vivid but two other positions were never encoded cleanly.

Separate encoding errors from comparison errors

A wrong answer can originate in two different phases. You may have stored the memory array inaccurately, or you may have stored it well but compared the probe inefficiently. The game cannot read your internal state, so use the pattern of errors and timing as clues. If you often feel unsure before the delay even starts, work on the study pass. If the memory feels clear but you hesitate only after the probe appears, use a more systematic comparison order.

One practical experiment is to replay the same deterministic seed twice. On the first run, use your normal method. On the second, keep the material identical but commit to a fixed scan order. If the second run reduces probe time without harming accuracy, the improvement likely came from comparison discipline rather than easier arrays.

Use set-size results as within-task evidence

The result cards are most useful as comparisons inside this game. If accuracy is similar at small and large set sizes but median response time increases, the extra load may be affecting comparison speed more than memory fidelity. If Change accuracy drops first, altered items may be slipping through as load rises. If Same accuracy drops first, uncertainty may be producing more false alarms.

Do not turn a single slope or one difficult block into a broad claim about memory ability. Session results are sensitive to attention, device input, interruptions and strategy. Look for repeatable patterns across several Clean blocks before changing your practice routine.

When to use Fresh block

Use Fresh block when the current concrete arrays are becoming familiar enough that recognition could replace active comparison. Use exact replay when you are testing a strategy. These two actions answer different questions: replay controls the material, while Fresh changes the material. Keeping that distinction clear makes local practice data easier to interpret.

What the result does not mean

DailyBrainArc Change Detection is a browser practice game. It does not diagnose a memory disorder, estimate IQ, provide a clinical capacity score, establish a population percentile, or promise improvement in school, work or health outcomes.