From 7739ec3b484a2de225959b8b1715e77d44ff4373 Mon Sep 17 00:00:00 2001 From: James Kent Date: Wed, 21 Jan 2026 09:49:58 -0600 Subject: [PATCH] wip: add new tutorial --- docs/tutorial/cognitive_control_aging.md | 218 +++++++++++++++++++++++ docs/tutorial/index.mdx | 27 ++- 2 files changed, 240 insertions(+), 5 deletions(-) create mode 100644 docs/tutorial/cognitive_control_aging.md diff --git a/docs/tutorial/cognitive_control_aging.md b/docs/tutorial/cognitive_control_aging.md new file mode 100644 index 0000000..01593e3 --- /dev/null +++ b/docs/tutorial/cognitive_control_aging.md @@ -0,0 +1,218 @@ +--- +sidebar_label: 'Manual Meta-Analysis Tutorial: Cognitive Control & Aging' +sidebar_position: 3 +--- + +# Cognitive Control & Aging + +*Neurosynth Compose* provides a streamlined workflow to help you curate studies, extract coordinate data, and run a cloud-based meta-analysis—while keeping your decision process transparent and reproducible. + +:::tip +This tutorial is written to feel similar to the [Manual Meta-Analysis tutorial (deprecated)](./manual.md), but showcases the new *Neurosynth Compose* interface and features. +::: + +## Your research question + +We’ll investigate *differences in cognitive control* between **older adults (65+)** and **younger adults**, using common cognitive control task families: + +- Stroop +- Flanker +- Go/NoGo +- Stop-signal +- N-back + +In practice, “cognitive control” is broad. A small up-front investment in defining what you mean (and what you do *not* mean) will save you from painful rework during Eligibility. + +:::note Figure placeholder +Add a simple schematic showing the study flow: Search → Screen → Eligibility → Extract → Specify → Run. +::: + +## Inclusion & exclusion criteria (copy/paste checklist) + +Write these down *before* you start screening. You’ll use them repeatedly. + +### Inclusion criteria + +1. **Older group is 65+** (older adults are defined as age 65 or above). +2. **Task is one of:** Stroop, Flanker, Go/NoGo, Stop-signal, N-back. +3. **Contrast is older vs younger** (a between-group comparison between an older and a younger group). +4. **Whole-brain results** (exclude ROI-only studies). + +### Exclusion criteria + +1. **Connectivity-only**: exclude studies that report only connectivity measurements (e.g., functional connectivity) without task-evoked coordinate peaks suitable for coordinate-based meta-analysis. +2. **No coordinates**: exclude studies that do not report any stereotaxic coordinates. + +:::tip +If you anticipate running separate meta-analyses by task (e.g., Stroop-only vs N-back-only), keep the same criteria and add task labels during Extraction using Annotations. +::: + +## Creating a new Project + +*Projects* contain all the Steps necessary to create a meta-analysis in Neurosynth Compose. + +To get started, *sign in*, and select **New Project**. +Name your project something you’ll recognize later, such as: “Cognitive control in older adults”. + +:::note Figure placeholder +Insert a screenshot of the new Project dialog and the three stages (Search & Curate → Extract & Annotate → Specify Meta-Analyses). +::: + +## Search & Curate (PRISMA workflow) + +The goal of this phase is to go from a *broad* literature search to a final, defensible StudySet that matches your criteria. + +Select the **PRISMA** curation workflow to keep your decisions structured (Identification → Screening → Eligibility → Included). + +### Importing studies (NeuroStore + PubMed) + +Click **Import Studies** to add papers to your PRISMA board. You can: + +- search within *NeuroStore* (fastest when the study already has extracted coordinates), or +- import from external sources like PubMed (best when you want maximal coverage). + +#### A practical PubMed query (starting point) + +Use this as a starting point, then refine based on what you see during Screening: + +``` +("fMRI" OR "PET" OR "neuroimaging" OR "functional MRI" OR "functional magnetic resonance imaging") +AND ("aging" OR "ageing" OR "older adult*" OR elderly) +AND (stroop OR flanker OR "go/no-go" OR go-nogo OR "stop signal" OR "stop-signal" OR "n-back" OR nback) +AND (young OR younger) +``` + +:::tip +If your query returns many electrophysiology-only papers, add a stricter imaging clause (e.g., require fMRI in the title/abstract). +::: + +#### NeuroStore search tips + +If you use NeuroStore search, try one task family at a time (e.g., “stroop older younger”), and filter to **coordinate** data when available. + +:::note Figure placeholder +Insert a screenshot of an Import Studies dialog with example search terms and filters for coordinate-based studies. +::: + +### Step 1: Identification (deduplicate) + +Import your candidate studies (via NeuroStore search, PubMed IDs, or your reference manager export), then remove duplicates during **Identification**. + +:::note Figure placeholder +Insert a screenshot of the Identification column showing duplicates and the deduplication UI. +::: + +### Step 2: Screening (title/abstract relevance) + +In **Screening**, you’re trying to answer one question quickly: + +“Is this plausibly a neuroimaging study of cognitive control that could include older (65+) and younger groups?” + +Helpful heuristics for Screening: + +- Look for the task keywords (stroop/flanker/go-no-go/stop-signal/n-back). +- Look for aging terms (older, elderly, aging, age-related). +- Look for imaging terms (fMRI, PET, neuroimaging). + +Don’t overthink Screening. If you’re unsure, **promote** it and decide during Eligibility with the full text. + +### Step 3: Eligibility (full-text criteria) + +Eligibility is where your criteria become enforceable. Open the full text and verify: + +- The older group is **65+** +- The task is **one of stroop/flanker/go-no-go/stop-signal/n-back** +- The results include a **between-group contrast** (older vs younger) +- The paper reports **stereotaxic coordinates** +- The paper is not **connectivity-only** + +:::tip +When you exclude a study at Eligibility, write the exclusion reason as if you’re explaining it to a collaborator who hasn’t read the paper. Future-you will thank you. +::: + +:::note Figure placeholder +Insert a screenshot of the Eligibility review dialog highlighting custom exclusion reasons (e.g., “Connectivity-only”, “No coordinates”, “No older-vs-younger contrast”). +::: + +### Step 4: Included (final check) + +Once you’ve promoted studies into **Included**, pause and sanity-check: + +- Do you have a reasonable mix of tasks, or is everything one task family? +- Are you accidentally including “aging” studies with no younger comparison group? +- Are any studies duplicates in disguise (e.g., same sample reported across multiple papers)? + +When ready, click *Move on to Extraction*. + +## Extract & Annotate + +The goal of this phase is to make sure the analyses you run are *exactly* the ones you intend: correct coordinates, correct task labels, and correct older-vs-younger contrasts. + +### Extract coordinates (when needed) + +For studies not already indexed with coordinates, add them from the paper: + +- Find the table(s) of peak coordinates (MNI or Talairach). +- Create an **Analysis** per relevant contrast. +- Enter each reported peak as a coordinate row (X, Y, Z), with the correct space. + +:::note Figure placeholder +Insert a screenshot of the Analysis editor showing the coordinate table and the coordinate-space selector. +::: + +:::caution +Many papers report multiple contrasts (and sometimes multiple tasks) within the same table. Be careful not to mix peaks from different contrasts into the same Analysis. +::: + +### Add Annotations to keep yourself organized + +Annotations let you tag each Analysis so you can filter precisely later. + +Recommended annotation columns for this tutorial: + +- `task` (String): `stroop`, `flanker`, `go_nogo`, `stop_signal`, `n_back` +- `contrast_direction` (String): `older_gt_younger` or `younger_gt_older` +- `eligible` (Boolean): a final, manual “yes/no” that reflects your criteria + +:::tip +Start with broad labels. You can always add finer-grained labels later (e.g., “Stroop interference” vs “congruent vs neutral”) once you see what the literature actually reports. +::: + +## Specify your meta-analysis (one question per model) + +From the Project page, proceed to **Meta-Analysis Specification**. + +To keep interpretations clean, define separate specifications when you’re asking meaningfully different questions, for example: + +- All tasks combined, `older_gt_younger` +- All tasks combined, `younger_gt_older` + +In the specification wizard, select the Annotation you want to filter on (e.g., `older_gt_younger`), then choose an algorithm/corrector with the default parameters unless you have a strong reason to change them. + +:::note Figure placeholder +Insert a screenshot of the meta-analysis specification wizard showing filtering by annotations. +::: + +## Run, inspect, and iterate + +Run your meta-analysis using the provided cloud workflow (e.g., the linked notebook). + +As you inspect results, keep the human question front-and-center: + +- “Do these maps reflect *cognitive control differences*—or did I unintentionally mix in working memory, motor inhibition, or error monitoring contrasts that aren’t comparable?” + +If something looks off, the most common fix is upstream: + +- tighten Eligibility criteria, +- correct mislabeled Analyses, +- split mixed Analyses into separate contrasts, +- or run task-specific sub-analyses. + +:::note Figure placeholder +Insert a screenshot of a results view (thresholded + unthresholded) and a short caption space for interpretation notes. +::: + +## Wrap-up + +You now have a reproducible workflow for a coordinate-based meta-analysis of **cognitive control and aging**, grounded in explicit criteria and organized around the contrasts you actually care about. + diff --git a/docs/tutorial/index.mdx b/docs/tutorial/index.mdx index 3410a36..281bef7 100644 --- a/docs/tutorial/index.mdx +++ b/docs/tutorial/index.mdx @@ -2,7 +2,7 @@ # Tutorials -import { Grid, Card, CardContent, Typography, Button } from '@mui/material'; +import { Card, CardContent, Typography, Button } from '@mui/material'; import { Link } from 'react-router-dom'; import DocCardList from '@theme/DocCardList'; @@ -15,8 +15,8 @@ from exploratory large-scale automated analyses to highly rigorous manual analys The choice of workflow depends on your research question and resources available for manual curation. We reccomend starting with the **manual meta-analysis** tutorial if you are new. -
- +
+ @@ -34,7 +34,7 @@ We reccomend starting with the **manual meta-analysis** tutorial if you are new. - + @@ -52,9 +52,26 @@ We reccomend starting with the **manual meta-analysis** tutorial if you are new. + + + + + Cognitive Control & Aging + + + Manual workflow with explicit criteria.

+
    +
  • Older adults defined as 65+
    • +
  • Tasks: Stroop, Flanker, Go/NoGo, Stop-signal, N-back
    • +
  • Between-group contrasts: older vs younger
    • +
  • Excludes connectivity-only and no-coordinate studies
    • +
+ + + +
### Advanced tutorials After you've completed the core tutorials above, you can continue your learning journey with [advanced tutorials](./tutorial/advanced). -