Guide
認知科學原理、模式解析與訓練策略
注意力並非單一能力，而是由數個子系統組成的認知架構。神經科學研究將注意力大致劃分為 選擇性注意力（從眾多刺激中鎖定目標）、 持續性注意力（長時間維持警覺）、 分配性注意力（同時處理多個任務）以及 執行控制（壓制無關干擾、靈活切換任務）。
日常生活中，這些能力影響著閱讀理解、工作效率、開車安全與人際溝通的品質。 研究顯示，以即時回饋搭配適當難度的練習，能夠在一定程度上優化注意力相關的處理效率。 Focus Training 的每個模式，都針對其中一至兩個面向設計了專屬的訓練情境。
大腦在搜尋目標時，會在「特徵搜尋」（Feature Search，目標具有獨特屬性，幾乎無需費力）與 「連結搜尋」（Conjunction Search，目標須靠多個特徵組合辨識，需逐一掃描）之間切換。 方格模式接近連結搜尋——數字在視覺上高度相似，必須逐格比對，因此能有效訓練搜尋速度與眼動效率。
訓練建議
不要從左上角開始逐格掃，改用「Z 字形」或「分區塊」策略，將格子切成四個象限輪流掃視，能大幅縮短平均搜尋時間。
心理學中的「多目標追蹤」（Multiple Object Tracking, MOT）是視覺注意力研究的經典範式。 在動態環境中同時追蹤多個移動物體，需要注意力在空間上持續分配，而非聚焦於單一位置。 球的數量直接對應認知負荷——9 顆球的難度遠低於 25 顆，因為工作記憶能同時追蹤的獨立物件有限（約 4±1 個）。
訓練建議
不要試圖追蹤所有球。先鎖定下一個目標球，其他球用周邊視覺監控即可。熟悉後再逐漸縮短「找到目標 → 點擊」的反應時間。
50 個數字分布在同心圓上，玩家需在完成一個數字後，立刻將注意力轉移到下一個目標區域。 這個過程結合了空間工作記憶（記住已找過的位置）與 抑制回返（Inhibition of Return）機制——大腦會短暫壓制剛剛注意過的位置， 避免重複搜尋同一區域，此機制在密集搜尋中扮演效率關鍵的角色。
訓練建議
先熟悉數字的大致分布規律（例如 1 永遠在圓心），建立空間地圖。找到規律後，眼睛可以提前移動到預期位置，而非找到後才移動。
干擾數字的出現模擬了真實情境中的「注意力競爭」——當環境中存在與目標高度相似的刺激時， 大腦需要投入更多的執行資源來篩選。後期的高密度干擾會顯著拉長反應時間， 這與心理學中的側翼效應（Flanker Effect）和注意力捕獲（Attentional Capture）現象一致。
訓練建議
不要掃視整個畫面，改為在螢幕中央等待目標出現。數字出現後再移動視線，而非提前搜尋，能減少被干擾數字誤導的機率。
Flanker 任務是認知科學中測量反應抑制的標準工具，最早由 Eriksen & Eriksen 於 1974 年提出。 「不一致」（Incongruent）試驗——側翼與目標方向相反——會產生反應競爭， 必須動用前額葉執行控制系統壓制自動反應，才能正確回答。 反應時間差（不一致試驗 − 一致試驗）稱為「干擾效應量」，是執行功能的直接指標。
訓練建議
視線鎖定中央箭頭，刻意忽略兩側。初期可以在心中默念「只看中間」來強化注意力定向，隨著練習次數增加，抑制會逐漸自動化。
工作記憶（Working Memory）是短暫保存並即時操作資訊的認知系統。 Baddeley 模型將其拆分為語音迴路、視覺空間模板（Visuospatial Sketchpad）與中央執行系統。 記憶模式主要訓練視覺空間模板——記憶數字在格子中的空間位置， 並在翻蓋後靠記憶重建完整的位置地圖。
訓練建議
不要逐個記憶位置，改用「群組化」策略：把 9 個格子想像成 3 個橫排，每排記「左中右分別是幾號」，能有效降低記憶負荷。
| 指標 | 代表意義 |
|---|---|
| 完成時間 | 整體處理速度與搜尋效率的綜合反映 |
| 錯誤次數 | 衝動控制與反應準確度；錯誤過多通常代表速度優先於準確 |
| 排名 PR | 相對於所有玩家的表現百分位，消除了「這個模式本來就快/慢」的基準差異 |
成績的提升在初期最為明顯（練習效應），之後會趨於穩定。 若單次成績與平均相差超過 20%，通常反映當下的疲勞、分心或環境干擾，而非真實能力的變化。
短期密集練習能產生顯著的任務內學習效果，但長時間連續遊玩會因注意力疲勞而降低表現。 建議每次訓練 10–15 分鐘，重複 2–3 次後休息，而非連續操作超過 30 分鐘。
Focus Training 定位為娛樂性互動工具， 遊戲成績不構成任何醫療、心理或認知能力的正式評估。若有相關需求，請諮詢專業醫療或心理機構。
Guide
Cognitive science principles, mode breakdowns, and training strategies
Attention is not a single faculty but a system of interconnected cognitive processes. Neuroscience broadly distinguishes selective attention (picking targets from competing stimuli), sustained attention (maintaining alertness over time), divided attention (handling multiple inputs simultaneously), and executive control (suppressing irrelevant responses and flexibly shifting focus).
These abilities affect everyday performance in reading comprehension, workplace productivity, driving safety, and communication quality. Research suggests that practice with real-time feedback and appropriately scaled difficulty can improve attentional processing efficiency. Each Focus Training mode is designed to target one or two of these dimensions in a controlled, measurable setting.
The brain switches between effortless feature search (when a target has a unique property) and demanding conjunction search (when the target must be identified by a combination of features requiring serial scanning). Grid mode closely resembles conjunction search — all numbers look alike, so each cell must be individually inspected. This directly trains search speed and oculomotor efficiency.
Training Tip
Avoid scanning from top-left. Instead, divide the grid into four quadrants and sweep each in turn using a Z-pattern. This structured scan strategy substantially reduces average search time.
Multiple Object Tracking (MOT) is a well-established paradigm in visual attention research. Tracking several moving objects simultaneously requires attention to be distributed spatially rather than focused on a single location. Capacity is typically limited to around 4 ± 1 independent objects, so the jump from 9 to 25 balls represents a qualitative shift in cognitive demand, not just a linear increase.
Training Tip
Don't attempt to track all balls. Lock onto the next target and monitor the rest with peripheral vision. As you improve, focus on compressing the time between "target located" and "click."
Clicking 50 numbers across expanding rings combines visuospatial working memory (remembering searched areas) with inhibition of return — a mechanism by which the brain briefly suppresses recently attended locations, preventing redundant re-scanning. In extended searches like this one, inhibition of return is a critical efficiency driver.
Training Tip
Learn the rough distribution of numbers across rings (e.g. 1 is always at the centre). Building an internal spatial map allows your eyes to anticipate the next location rather than reacting after the fact.
The growing field of distractors replicates real-world attention competition — when stimuli closely resemble the target, the executive system must invest more resources to filter them out. This mirrors attentional capture and flanker interference phenomena studied in cognitive psychology, where salient or similar distractors impair target detection speed.
Training Tip
Resist scanning the whole screen. Station your gaze near the centre and wait for the target to appear before moving. This reduces exposure to distractors and the risk of involuntary attentional capture.
The Flanker task, introduced by Eriksen & Eriksen in 1974, is a standard laboratory measure of response inhibition. Incongruent trials — where flankers oppose the target direction — generate response competition that recruits prefrontal executive control to suppress the prepotent incorrect response. The reaction time difference between incongruent and congruent trials (the "interference effect") is a direct index of executive function efficiency.
Training Tip
Fix your gaze on the centre arrow and actively ignore the flankers. Silently repeating "centre only" during early practice reinforces intentional attentional direction. With repetition, suppression becomes increasingly automatic.
Working memory is the cognitive system that temporarily stores and manipulates information in real time. Baddeley's model divides it into a phonological loop, a visuospatial sketchpad, and a central executive. Memory mode targets the sketchpad directly: encoding the spatial positions of nine numbers and reconstructing the map from memory once the cards are flipped face-down.
Training Tip
Avoid memorising positions one by one. Instead, chunk the grid into three rows and encode each row as "left / centre / right = X / Y / Z." Chunking reduces memory load from nine independent items to three grouped patterns.
| Metric | What It Reflects |
|---|---|
| Completion time | Overall processing speed and search efficiency |
| Error count | Impulse control and response accuracy — frequent errors indicate speed prioritised over precision |
| PR (percentile rank) | Performance relative to all players, removing mode-specific baseline differences |
Improvement is most rapid in the early sessions (practice effect) and then plateaus. Single-session scores that deviate more than 20% from your average typically reflect fatigue, distraction, or environmental noise rather than genuine ability change.
Short, focused sessions produce stronger within-task learning than extended continuous play. Ten to fifteen minutes of active play, repeated two or three times with breaks in between, is more effective than a single uninterrupted thirty-minute session.
Focus Training is an entertainment tool. Scores reflect performance under specific conditions and do not constitute a medical, psychological, or clinical cognitive assessment. For any evaluation needs, please consult a qualified professional.