Brain Exposome

腦健康 ExposomeBrain-Health Exposome 檢測Check-in

根據 2024 年《Lancet》失智症委員會的 14 項可調整風險因子，以及你一生中居住地的空氣污染（PM2.5）累積暴露，用淺白的方式看看這些「exposome（環境暴露總和）」與腦健康的關係，並給你可行動的方向。Using the 14 modifiable risk factors from the 2024 Lancet Commission on dementia, plus the cumulative air-pollution (PM2.5) exposure across the places you've lived, see in plain terms how your lifelong 'exposome' relates to brain health — and what you can act on.

🔒 所有計算都在你的瀏覽器內完成，資料不會上傳或儲存到任何伺服器。All calculations run entirely in your browser — nothing is uploaded or stored on any server.

⚠️ 重要提醒：⚠️ Important: 本工具僅供一般健康教育與自我覺察之用，並非醫療器材、不是診斷或篩檢工具，無法預測任何人是否會罹患失智症。This tool is for general health education and self-awareness only. It is not a medical device, not a diagnostic or screening test, and cannot predict whether anyone will develop dementia.

  • 結果描述的是「族群層級」的關聯（來自已發表研究），而不是對「你個人」的預測；同樣的因子，每個人的結果差異其實很大。Results describe population-level associations (from published research), not a prediction about you personally — individual outcomes vary a great deal.
  • 這裡的「腦齡加速」是一個教育性的比喻數字，不是用腦部影像（MRI／EEG）量出來的真實腦齡。The 'brain-age acceleration' here is an educational illustration, not a real brain age measured from imaging (MRI/EEG).
  • 本工具不能取代醫療專業判斷。如有任何健康疑慮，請諮詢合格的醫療人員。This does not replace professional medical advice. Consult a qualified healthcare provider for any health concern.
  • PM2.5 採用 ACAG 衛星推估的逐年（1998–2024）縣市資料（1998 年以前的居住年份沿用 1998 值）。地圖上的失智盛行率為縣市層級的模型估計值，非實測。PM2.5 uses ACAG satellite-derived annual (1998–2024) county data (residence years before 1998 reuse the 1998 value). The dementia prevalence on the map is a county-level modelled estimate, not measured.
  • 為了少填一步，第一段居住地的「國家」會依你目前的網路所在地或瀏覽器語系自動預設（不經第三方、不會索取定位權限），你可以隨時自行更改。所有計算都在你的瀏覽器完成，不會上傳。To save a step, the first residence's country is pre-filled from your current network location or browser locale (no third party, no location-permission prompt); you can change it anytime. All calculations run in your browser and nothing is uploaded.
  1. 1基本資料About you
  2. 2居住史Places
  3. 3健康與生活Health
  4. 4結果Results

1. 關於你About you

2. 你住過的地方Places you've lived

加入你人生各階段主要居住的縣市與大約年份。系統會用該地區的年均 PM2.5 估算你「時間加權」的長期空氣污染暴露。年份可粗估即可。Add the main counties you've lived in across life, with rough years. We estimate your time-weighted long-term PM2.5 exposure from each area's annual mean. Approximate years are fine.

國家Country 地區Region 從From 到To PM2.5

3. 健康與生活型態Health & lifestyle

不確定就選「不確定」，不會被算成風險。這些對應到 Lancet 委員會可調整的失智風險因子。Pick 'Not sure' if you don't know — it won't be counted as a risk. These map to the Lancet Commission's modifiable dementia risk factors.

👂 感官👂 Sensory

❤️ 心血管代謝❤️ Cardiometabolic

🧠 心理健康🧠 Mental health

🤝 社會與認知🤝 Social & cognitive

🚬 生活型態與外傷🚬 Lifestyle & injury

4. 你的結果Your results

教育性「腦齡加速」估計Educational 'brain-age acceleration' estimate
—

各面向風險換算腦齡老化概況Brain-aging by domain (risk-converted)

你的風險因子：說明與建議Your flagged factors: what they mean & what to do

    計算方式How this is calculated

    每個「有」的因子會帶入已發表的相對風險（RR，主要來自 2024 年 Lancet 委員會與其引用的統合分析）。因為這些因子彼此相關、會重複計算，我們在合併時乘上一個縮減係數 λ = 0.6（近似委員會的 communality 調整），並設上限，避免高估。合併後的有效 RR 再換算成「腦齡加速」年數：Δ年 ≈ log₂(RR) × 5.5（失智風險大約每 5–6 年翻倍）。空氣污染以連續劑量處理（每 +5 µg/m³ 約 HR 1.08），因此不再重複計入委員會的類別式空污因子。Each 'yes' factor contributes a published relative risk (RR, mainly from the 2024 Lancet Commission and the meta-analyses it cites). Because these factors are correlated and would double-count, we combine them with a shrinkage factor λ = 0.6 (approximating the Commission's communality adjustment) and cap the total to avoid overstating. The combined effective RR is converted to 'brain-age acceleration' years: Δyears ≈ log₂(RR) × 5.5 (dementia risk roughly doubles every 5–6 years). PM2.5 is handled continuously (≈ HR 1.08 per +5 µg/m³), so the Commission's categorical air-pollution factor is not added on top.

    假設與限制：λ = 0.6 與腦齡年數上限是本工具為避免高估所設的參數（非已發表常數）；PM2.5 採 ACAG 衛星年均資料（依居住縣市／國家與年份對應），歷史不足年份以最接近年份外推。這是教育性估計，非個人預測。Assumptions: λ = 0.6 and the brain-age cap are this tool's parameters to avoid overstatement (not published constants); PM2.5 uses ACAG satellite annual data (matched to your county/country and year), with missing historical years extrapolated from the nearest available year. This is an educational estimate, not a personal prediction.

    全球地圖：高齡化、認知退化與危險因子概況Global map: aging, cognitive decline & risk factors

    預設是「🌍 全球總覽」地球儀——每個國家以近年 PM2.5 或可調控失智風險（PAF）著色，色階以 WHO 標準（≤5 µg/m³）為基準、可跨國比較。點任一國（或用選單切換到 20 多國）即可下鑽到該國各行政區的細節地圖，看更細的 PM2.5 與失智盛行「模型估計值」。把游標移到區域上看數值。The default is a 🌍 global overview globe — each country shaded by recent PM2.5 or modifiable dementia risk (PAF), on a WHO-anchored scale (≤5 µg/m³) comparable across countries. Click a country (or pick one of 20+ from the dropdown) to drill into its admin-1 detail map, with finer PM2.5 and MODELLED dementia-prevalence estimates. Hover a district for values.

    ※ 台灣失智盛行率為鄉鎮市區層級「65 歲以上盛行率」的模型估計（NHRI 年齡別盛行率 × 內政部 #77132 各區單一年齡人口，見下方說明表），非實測。PM2.5 為 ACAG 衛星資料（V6.GL.03）；台灣以外為各行政區（Natural Earth 界線）的衛星網格平均值。美國為本土 48 州＋DC（不含阿拉斯加／夏威夷）。※ Taiwan dementia prevalence is a township-level modelled estimate of prevalence among residents aged 65+ (NHRI age-band rates × MOI #77132 single-year-age population per township — see the method table below), not measured. PM2.5 is ACAG satellite data (V6.GL.03); outside Taiwan, values are grid means per admin-1 unit (Natural Earth boundaries). USA shows the contiguous 48 states + DC (Alaska/Hawaii excluded).

    資料齊全度與來源Data coverage & sources

    逐一揭露本工具各類資料的齊全度與出處：切換上方資料類型，表列各國的國家級／省縣級資料來源與年份。狀態：✅ 已用於工具・◐ 部分（種子估計）・○ 已找到來源待接入・— 尚無資料。皆為模型估計／最新可得估計，會持續補齊。Full disclosure of each data layer's coverage and provenance: pick a data type; the table lists every country's national / sub-national source and year. Status: ✅ in the tool · ◐ partial (seed) · ○ source identified, not yet wired · — none yet. All modelled / latest-available estimates; being filled in over time.

    資料來源與參考文獻Sources & references

    本工具整合以下公開文獻、政府與國際組織資料庫。所有失智盛行率與腦齡數字皆為教育性模型估計，非醫療診斷。This tool builds on the public literature, government, and international datasets below. All dementia-prevalence and brain-age figures are educational modelled estimates, not medical diagnoses.

    ※ 資料庫類來源（GBD、內政部戶政司 #77132、日本 e-Stat、韓國 KOSIS、ACAG）均於 2026-07-14 存取；各條標註所用資料年份。※ Database sources (GBD, MOI #77132, Japan e-Stat, Korea KOSIS, ACAG) were accessed 2026-07-14; each entry notes the data year used.

    ① 失智風險與腦齡模型① Dementia risk & brain-age model

    • Livingston G, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet 2024;404:572–628. doi:10.1016/S0140-6736(24)01296-0
    • Moguilner S, Ibáñez A, et al. Brain clocks capture diversity and disparities in aging and dementia. Nat Med 2024;30:3646–57. doi:10.1038/s41591-024-03209-x
    • Long-term air pollution exposure and incident dementia: a systematic review and meta-analysis. Lancet Planet Health 2025 (≈ HR 1.08 per 5 µg/m³ PM2.5). doi:10.1016/S2542-5196(25)00118-4
    • Ziegler-Graham K, et al. Worldwide variation in the doubling time of Alzheimer's disease incidence rates (≈ 5.5 years). Alzheimers Dement 2008;4:316–23. doi:10.1016/j.jalz.2008.05.2479

    ② 失智盛行率調查與資料庫② Dementia-prevalence surveys & databases

    • 台灣：國家衛生研究院（NHRI）112年(2023)全國社區失智症流行病學調查（年齡別盛行率）；年齡別盛行率方法參 Sun Y, et al. PLoS ONE 2014。Taiwan: NHRI 2023 nationwide community dementia survey (age-band prevalence); method per Sun Y, et al. PLoS ONE 2014. NHRI · doi:10.1371/journal.pone.0100303
    • 美／巴／墨（州級）與日／韓（全國費率）失智盛行率之骨幹：Rate backbone for US/BR/MX (sub-national) and JP/KR (national rates): IHME Global Burden of Disease GBD 2023 — Alzheimer's disease & other dementias, prevalence（資料年份 2023data year 2023）. vizhub.healthdata.org/gbd-results
    • 中國：China: Jia L, et al. Prevalence of dementia in China (CCAS). Lancet Public Health 2020. doi:10.1016/S2468-2667(20)30185-7 · Liu, Gao, et al. Geographical variation in dementia prevalence across China (per-province). Lancet Reg Health – West Pac 2024. PMC11225804
    • 印度：India: Lee J, et al. Prevalence of dementia in India: national and state estimates (LASI-DAD). Alzheimer's & Dementia 2023. doi:10.1002/alz.12928
    • 日本：Japan: Ninomiya T, et al. 久山町研究（37 年趨勢，年齡別盛行率）Hisayama Study (37-year trends, age-specific prevalence). Alzheimer's Research & Therapy 2025. doi:10.1186/s13195-025-01909-1 · 厚生労働省 全國認知症・MCI 高齡者推計（Ninomiya 研究班，2024）MHLW nationwide dementia & MCI estimates (Ninomiya group, 2024) mhlw.go.jp
    • 韓國：Korea: 중앙치매센터 國立失智症中心《大韓民國失智現況》（年齡別유병률・MCI）National Institute of Dementia — Korean Dementia Observatory (age-specific prevalence & MCI) nid.or.kr
    • MCI（輕度認知障礙）圖層 — 底色為區域統合估計：Bai et al. 2022《Age & Ageing》依世界銀行區域的 MCI 盛行率（50 歲以上、社區）。約 70 國以其全國研究值覆寫底色（限合理範圍 3–40%，超出者多為 MoCA 篩檢高估、退回區域值），例：台灣 Sun 2014/2017、日本 MHLW 2024、南韓 KDO 2020、印度 LASI-DAD、中國 Xue 2021；逐國來源見 scripts/mci-scd-sources.json。定義／年齡層不一，不可跨國比較，為模型估計。MCI (mild cognitive impairment) layer — baseline is a regional pooled estimate: Bai et al. 2022 (Age & Ageing), MCI prevalence by World Bank region (adults 50+, community). ~70 countries overwrite the baseline with a national-study value (plausible 3–40%; higher MoCA screening figures revert to regional), e.g. Taiwan Sun 2014/2017, Japan MHLW 2024, Korea KDO 2020, India LASI-DAD, China Xue 2021; per-country sources in scripts/mci-scd-sources.json. Definitions/age bands vary — not comparable, a modelled estimate. Bai 2022 · MHLW
    • MCI 佐證：多篇全球統合估計一致支持上列區域底色 — Song 2023 全球 19.7%、Salari 2025 老年族群 23.7%、拉丁美洲專屬統合 14.95%（2021）、中國 Lu 2021 12.2% / Xue 2021 15.4%（差異源於年齡切點與量表）。SCD（主觀認知衰退）已上獨立「SCD（自述）」圖層（38 國，多為單題自述、各自尺度、不可跨國比較，7–76%）；仍無全球統合可當底色，逐國來源見 scripts/mci-scd-sources.json。MCI corroboration: several global meta-analyses agree with the regional baseline above — Song 2023 global 19.7%, Salari 2025 geriatric 23.7%, Latin America pooled 14.95% (2021), China Lu 2021 12.2% / Xue 2021 15.4% (differences reflect age cut-off and instrument). SCD (subjective cognitive decline) now has its own layer (38 countries, mostly single-item self-report, own scale, not comparable, 7–76%); still no global meta as a baseline; per-country sources in scripts/mci-scd-sources.json. Song 2023 · Salari 2025

    ③ 空氣污染（PM2.5）③ Air pollution (PM2.5)

    • Atmospheric Composition Analysis Group (ACAG), Washington University in St. Louis — surface PM2.5 V6.GL.03 (satellite-derived, annual 1998–2024), CC BY 4.0. Methods: van Donkelaar A, et al. Environ Sci Technol 2021 (doi:10.1021/acs.est.1c05309); Shen S, et al. ACS ES&T Air 2024 (doi:10.1021/acsestair.3c00054). dataset
    • WHO 全球空氣品質指引 2021（年均 PM2.5 建議 ≤ 5 µg/m³）。WHO Global Air Quality Guidelines 2021 (recommended annual PM2.5 ≤ 5 µg/m³). who.int

    ④ 地理界線④ Administrative boundaries

    • 台灣：內政部國土測繪中心（NLSC）鄉鎮市區界線，經 taiwan-atlas 轉製（OGDL-Taiwan-1.0）。Taiwan: township boundaries from the National Land Surveying and Mapping Center (NLSC), via taiwan-atlas (OGDL-Taiwan-1.0). data.gov.tw/7441
    • 其他國家：Other countries: Natural Earth admin-1 boundaries (public domain). naturalearthdata.com

    ⑤ 人口與國際報告⑤ Population & international reports

    • 台灣人口：內政部戶政司 #77132 鄉鎮市區單一年齡人口（資料月份 2026-06）。Taiwan population: MOI #77132 township single-year-age population (data 2026-06). data.gov.tw/77132
    • 日本人口：總務省統計局 e-Stat 都道府県・年齡別人口（資料 2024）。韓國人口：統計廳 KOSIS 시도・연령별 인구（資料 2026）。Japan population: Statistics Bureau e-Stat, prefecture × age (data 2024). Korea population: Statistics Korea KOSIS, Si-do × age (data 2026). e-stat.go.jp · kosis.kr
    • 其餘各國 admin-1 人口分母：WorldPop 2020 1km 年齡／性別網格（資料 2020，CC BY 4.0）；UN World Population Prospects 為國家別校準。部分高所得國家 WorldPop 年齡結構在區間近乎一致，故以全國估計值呈現。Admin-1 population denominator for the remaining countries: WorldPop 2020 1km age/sex grids (data 2020, CC BY 4.0); UN World Population Prospects for national calibration. For some high-income countries WorldPop's age structure is near-uniform across areas, so a single national estimate is shown. worldpop.org · population.un.org/wpp
    • WHO. 失智症全球公共衛生應對現況報告，2021（全球失智人口 55M → 78M（2030）→ 139M（2050））。Global status report on the public health response to dementia, 2021 (55M now → 78M by 2030 → 139M by 2050). who.int
    • 高齡化社會分級標準（高齡化 ≥7%、高齡 ≥14%、超高齡 ≥20%，以 65 歲以上人口占比計）：聯合國經濟社會事務部《World Population Ageing》；7% 門檻源自 UN 1956《The Aging of Populations and Its Economic and Social Implications》Population Studies No. 26。全球高齡人口占比採 World Bank SP.POP.65UP.TO.ZS（2025，存取 2026-07-16）；台灣不在該庫，改採內政部戶政司 2025 年底 20.06%。Aging-society tiers (aging ≥7%, aged ≥14%, super-aged ≥20%, by share of population aged 65+): UN DESA, World Population Ageing; the 7% threshold originates in UN (1956), The Aging of Populations and Its Economic and Social Implications, Population Studies No. 26. Global 65+ shares use World Bank SP.POP.65UP.TO.ZS (2025, accessed 2026-07-16); Taiwan, excluded from that database, uses MOI 2025 year-end (20.06%). un.org/ageing · data.worldbank.org

    ⑥ 可調控風險因子與 PAF（風險圖層）⑥ Modifiable risk factors & PAF (risk layers)

    • 相對風險（RR）與可歸因比例（PAF）架構取自 Livingston 2024 Lancet 委員會（見 ①）。PAF = P(RR−1)/[1+P(RR−1)]；5 因子合併採 1−Π(1−PAF)（假設獨立），為下限估計。Relative risks (RR) and the population-attributable-fraction (PAF) framework are from the Livingston 2024 Lancet Commission (see ①). PAF = P(RR−1)/[1+P(RR−1)]; the 5 factors are combined via 1−Π(1−PAF) (independence-assumed), a floor estimate.
    • 高血壓／糖尿病／肥胖（全球各國、全成人、年齡標準化、實測，含台灣）：Hypertension / diabetes / obesity (all countries, adult, age-standardised, measured; incl. Taiwan): NCD Risk Factor Collaboration (NCD-RisC) — Lancet 2017;389:37–55 doi（血壓，資料 2015BP, data 2015）· Lancet 2024;404:2077–93 doi（糖尿病，資料 2022diabetes, data 2022）· Nature 2026（BMI≥30，資料 2024data 2024）ncdrisc.org
    • 吸菸與身體活動不足（全球各國，台灣除外見下；年齡標準化，吸菸資料至 2025、身體活動至 2022，存取日 2026-07-16）：Smoking & insufficient physical activity (all countries except Taiwan, see below; age-standardised, smoking to 2025, activity to 2022, accessed 2026-07-16): WHO Global Health Observatory（指標indicators M_Est_tob_curr, NCD_PAC）. who.int/data/gho
    • 台灣吸菸與身體活動不足：2021 年國民健康訪問調查（衛福部國民健康署）。WHO GHO 不含台灣，故由此填補；台灣高血壓／糖尿病／肥胖仍採 NCD-RisC 台灣值以與各國可比。Taiwan smoking & inactivity: 2021 National Health Interview Survey (Taiwan Health Promotion Administration). WHO GHO has no Taiwan row, so these fill it; Taiwan's hypertension/diabetes/obesity use NCD-RisC for cross-country comparability. hpa.gov.tw