Imported from bg-szy/TOP-SKILLS (skills/awesome-skills/scientific-brainstorming/SKILL.md). Install upstream with npx skills add bg-szy/TOP-SKILLS --skill scientific-brainstorming. Copyright stays with the author (CC-BY-4.0).

Scientific Brainstorming

Overview

Scientific brainstorming is a structured ideation process for generating, connecting, and evaluating research ideas. Unlike casual brainstorming, scientific brainstorming applies formal methodologies (SCAMPER, TRIZ, Morphological Analysis, etc.) matched to the specific creative challenge. The process moves through divergent exploration, connection-making, critical evaluation, and synthesis to produce actionable research directions with testable hypotheses.

Key Concepts

Core Principles

Five principles guide effective scientific brainstorming sessions:

1. Collaborative: Brainstorming works best as dialogue, not monologue. Build on each other's ideas rather than presenting finished thoughts. Use "Yes, and..." framing to extend ideas before evaluating them. In AI-assisted sessions, the scientist contributes domain expertise and the AI contributes breadth and pattern-matching across disciplines.

2. Curious: Approach the problem space with genuine curiosity. Ask "what if" and "why not" before "why." Suspend expertise-driven assumptions temporarily to allow unexpected connections. Experts often dismiss novel directions because they conflict with established mental models -- curiosity counteracts this.

3. Domain-Aware: Ground brainstorming in real scientific constraints. Ideas must eventually connect to testable hypotheses, available methods, and feasible experiments. Domain knowledge channels creativity productively. Pure creativity without domain grounding produces ideas that cannot be tested; pure domain expertise without creativity produces incremental work.

4. Structured: Use formal ideation methods rather than unguided free association. Structure prevents cognitive fixation (repeatedly returning to the same idea space) and ensures systematic coverage of the possibility space. Unstructured brainstorming sessions typically explore less than 20% of the available idea space.

5. Challenging: Actively seek ideas that feel uncomfortable or counterintuitive. The most productive brainstorming sessions push past obvious solutions into territory that requires deeper analysis. If every idea generated feels reasonable and safe, the session is not pushing hard enough.

Brainstorming Methods Catalog

Nine structured methodologies, each suited to different creative challenges:

SCAMPER Operators in Detail

The seven SCAMPER operators applied to scientific contexts:

• Substitute: Replace one component with another. What material, reagent, model organism, or technique could replace the current one? Example: replace mouse models with organoids; substitute CRISPR for siRNA knockdown.

• Combine: Merge two approaches, techniques, or datasets. What happens if you combine two assays into one? Two datasets from different modalities? Example: combine proteomics and metabolomics in a single sample preparation.

• Adapt: Borrow a technique from another field. What methods from physics, engineering, or computer science could solve this biological problem? Example: adapt semiconductor lithography for tissue engineering scaffolds.

• Modify: Change the scale, frequency, intensity, or duration. What if you ran the experiment 10x faster, at 100x concentration, or at a different temperature? Example: single-molecule resolution instead of bulk measurement.

• Put to other use: Apply an existing tool or finding to a different purpose. What other questions could this dataset answer? What other diseases could this drug treat? Example: repurpose failed drug candidates for new indications.

• Eliminate: Remove a step, component, or constraint. What if you eliminated the purification step? The control group? The assumption of linearity? Example: label-free detection instead of fluorescent tagging.

• Reverse: Invert the order, direction, or perspective. What if you worked backwards from the output? Reversed the cause-effect relationship? Example: start from the phenotype and work backwards to the genotype.

TRIZ Core Concepts

TRIZ (Theory of Inventive Problem Solving) provides three key concepts for scientific brainstorming:

• Technical Contradiction: Improving one parameter worsens another. Example: increasing drug selectivity (desired) reduces potency (undesired). TRIZ provides 40 inventive principles organized in a contradiction matrix to resolve such trade-offs systematically rather than by trial and error.

• Ideal Final Result (IFR): Envision the perfect solution where the desired function is achieved with zero cost, zero harm, and zero complexity. Working backwards from the IFR reveals which constraints are real and which are assumed. Example: the ideal drug delivers itself precisely to the target, requires no administration, and has no side effects -- what existing technology gets closest?

• Inventive Principles: The most commonly applicable principles in scientific research include:

  • Segmentation: divide a system into independent parts
  • Extraction: separate an interfering component or property
  • Local Quality: vary conditions locally rather than globally
  • Asymmetry: introduce useful asymmetry into a symmetric system
  • Nesting: place one system inside another (e.g., nanoparticle within liposome)
  • Prior Action: perform required changes in advance (e.g., pre-functionalization)
  • Dynamization: allow a system to change to achieve optimal conditions at each stage

Biomimicry Process Steps

The Biomimicry design spiral follows five steps:

1. Define: What function do you need? Frame the problem as a verb, not a noun. "How does nature filter?" not "How does nature make a filter?"
2. Biologize: Translate the function into biological terms. "What organisms need to separate particles from fluid?"
3. Discover: Search for organisms that have solved this problem. Use resources like AskNature.org or consult biological literature.
4. Abstract: Extract the design principle from the biological model. Not "copy the kidney" but "use countercurrent flow with selective permeability membranes."
5. Apply: Translate the abstracted principle back to the research problem. Design the experiment or technology using the biological principle as inspiration.

Method Categories

Brainstorming methods serve three distinct cognitive functions. Effective sessions use methods from multiple categories:

• Divergent methods expand the idea space: SCAMPER, Provocation, Random Input, Future Backwards. Use these when you need more options or feel stuck with too few ideas. These methods deliberately break existing thought patterns.
• Connecting methods find relationships between ideas: Morphological Analysis, Biomimicry, Reverse Assumptions. Use these when you have elements that might relate but have not identified how. These methods build structure across disparate concepts.
• Convergent methods evaluate and select: Six Thinking Hats, TRIZ. Use these when you have many ideas and need to identify which are most promising or how to resolve contradictions. These methods apply systematic judgment.

A complete brainstorming session should use at least one method from each category, typically in the order: divergent first, connecting second, convergent third.

Method Combinations

Certain methods pair well for deeper exploration. The key principle is to combine methods from different cognitive categories (divergent + connecting, or connecting + convergent):

• SCAMPER + Six Hats: Generate modifications with SCAMPER (divergent), then evaluate each modification from six perspectives (convergent). Particularly effective for experimental protocol refinement.

• Morphological Analysis + TRIZ: Map the design space with Morphological Analysis (connecting), then use TRIZ (convergent) to resolve contradictions that emerge in promising but conflicting combinations.

• Biomimicry + Provocation: Use Biomimicry (connecting) to find natural solutions, then apply Provocation (divergent) to push beyond biological constraints into engineered solutions that nature has not explored.

• Reverse Assumptions + Future Backwards: Challenge current assumptions first (connecting), then project forward from those reversed assumptions to envision alternative research trajectories (divergent).

• Random Input + Morphological Analysis: Use Random Input (divergent) to discover a new dimension for the morphological matrix (connecting) that was not previously considered. This often reveals blind spots in the design space.

• Three-method sequence (recommended for full sessions): Start with a divergent method (SCAMPER or Provocation, 15 min), then a connecting method (Morphological Analysis or Biomimicry, 15 min), then a convergent method (Six Hats or TRIZ, 15 min). This ensures all cognitive modes are exercised.

Decision Framework

Select a brainstorming method based on your primary creative challenge:

What is your brainstorming challenge?
├── Stuck / no new ideas coming
│   ├── Need a completely fresh perspective → Provocation Technique
│   └── Need external stimulus → Random Input
├── Improving an existing method or system
│   ├── Incremental improvements → SCAMPER
│   └── Hit a technical contradiction → TRIZ
├── Exploring a design space systematically
│   ├── Known dimensions, many options → Morphological Analysis
│   └── Unknown dimensions, need inspiration → Biomimicry
├── Need multiple perspectives on a decision
│   └── → Six Thinking Hats
├── Questioning fundamental assumptions
│   └── → Reverse Assumptions
└── Planning long-term research direction
    └── → Future Backwards

Best Practices

1. Start with context understanding before diverging: Spend adequate time understanding the problem space, existing literature, and constraints before generating ideas. Ideas generated without context tend to reinvent existing solutions or violate known constraints. Allocate at least 20% of session time to context mapping.

2. Use "Yes, and..." to build on ideas: When an idea is proposed, extend it before evaluating it. "Yes, and we could also..." keeps the creative momentum. Premature "but" statements shut down exploration before the idea space is fully mapped.

3. Balance divergent and convergent phases: Alternate between generating ideas (divergent) and evaluating them (convergent). Do not mix the two simultaneously -- generating and judging at the same time reduces both quantity and quality of ideas. Explicitly signal phase transitions: "We are now switching from idea generation to evaluation."

4. Cross-domain analogies unlock novel connections: The most innovative scientific ideas often come from applying principles from one domain to another. Deliberately seek analogies from biology, engineering, physics, mathematics, or even social systems. Biomimicry and Random Input methods formalize this practice. Keep a personal catalog of interesting mechanisms from outside your field.

5. Combine methods for deeper exploration: No single method covers all cognitive modes. Use a divergent method (SCAMPER, Provocation) followed by a connecting method (Morphological Analysis) followed by a convergent method (Six Hats, TRIZ). Three-method sequences produce richer results than single-method sessions.

6. Document ideas immediately: Capture every idea during the session, even ones that seem weak. Ideas that appear marginal during brainstorming often become valuable when revisited with fresh perspective. Use structured formats (idea + rationale + potential test) rather than bare lists. A structured record also prevents repeating the same brainstorming ground in future sessions.

7. Know when to switch methods: If a method stops producing new ideas after 10-15 minutes, switch to a different method rather than forcing it. Method fatigue is real -- the same cognitive frame produces diminishing returns. Signs to switch: repeated similar ideas, long silences, circling back to previously stated ideas, or frustration with the method's constraints.

8. End with synthesis, not just a list: Every brainstorming session should conclude with clustering related ideas, identifying the top 3-5 most promising directions, and defining concrete next steps (literature search, feasibility check, preliminary experiment). An unsynthesized idea list rarely leads to action.

Common Pitfalls

1. Evaluating too early: Critiquing ideas during the divergent phase kills creative momentum. Participants self-censor to avoid criticism, reducing both the quantity and novelty of ideas generated.

   • How to avoid: Explicitly separate divergent (generation) and convergent (evaluation) phases. During divergent phases, the only permitted response is "Yes, and..." or "What if we also..." Record all ideas without judgment.

2. Staying in the comfort zone: Defaulting to familiar methods and familiar idea spaces. Scientists tend to brainstorm within their own domain using their established mental models, which produces incremental rather than transformative ideas.

   • How to avoid: Deliberately use at least one unfamiliar method per session. If you always use SCAMPER, try Biomimicry or Provocation. Include at least one cross-domain analogy in every session. Read outside your field regularly.

3. Skipping the context phase: Jumping directly into idea generation without understanding the problem space, existing solutions, and constraints. This wastes time rediscovering known solutions and produces ideas that have already been tried.

   • How to avoid: Always begin with Phase 1 (Understand Context) from the Workflow below. Summarize the current state of knowledge, key constraints, and what has already been tried before generating new ideas.

4. Forcing a method when it is not working: Persisting with a method that is not generating results because of sunk-cost fallacy or belief that the method should work for this problem type.

   • How to avoid: Set a time limit (15-20 minutes) per method. If ideas are not flowing, switch methods without guilt. The Decision Framework above helps select an alternative method based on the specific challenge type.

5. Focusing on quantity without connection-making: Generating a long list of disconnected ideas without identifying patterns, themes, or combinations. A list of 50 unrelated ideas is less useful than 15 ideas organized into 3 coherent research directions.

   • How to avoid: After each divergent burst, spend time on connection-making (Phase 3 in the Workflow). Use affinity mapping: group related ideas, name the clusters, look for bridges between clusters.

6. Treating brainstorming as a monologue: One person (or one AI) generating ideas while others passively listen. This misses the combinatorial power of multiple perspectives interacting and building on each other.

   • How to avoid: Structure the session for dialogue. Alternate who proposes and who extends. Use Six Thinking Hats to ensure everyone contributes from different angles. In AI-assisted brainstorming, provide your own ideas and domain expertise rather than only receiving suggestions.

7. No follow-through after the session: Generating exciting ideas but never converting them into concrete research actions. Brainstorming without follow-through is intellectual entertainment, not research methodology.

   • How to avoid: Always complete Phase 5 (Synthesis and Next Steps). Assign specific follow-up actions (literature search, feasibility analysis, preliminary experiment design) with deadlines. Revisit the brainstorming output within one week.

Workflow

Phase 1: Understand Context (15-20% of session time)

1. Map the problem space: Define the research question, current state of knowledge, key constraints (time, budget, equipment, expertise), and what solutions have already been attempted. Identify the specific gap or challenge that brainstorming aims to address.
2. Review prior work: Briefly summarize what is already known. What approaches have been tried? What failed and why? What are the current leading hypotheses? This prevents reinventing the wheel during divergent exploration.
3. Identify the creative challenge type: Is the problem about being stuck, improving something existing, exploring systematically, resolving contradictions, or something else? Use the Decision Framework to classify.
4. Select initial method(s): Choose 1-2 brainstorming methods based on the creative challenge type. Have a backup method ready in case the first does not produce results.
   • Transition criterion: Proceed to Phase 2 when the problem is clearly defined, the knowledge landscape is mapped, and at least one method is selected.

Phase 2: Divergent Exploration (30-40% of session time)

1. Apply the selected method systematically: Work through the chosen brainstorming method completely. For SCAMPER, apply all 7 operators to the problem. For Morphological Analysis, define all dimensions before exploring combinations. For Biomimicry, complete the full define-biologize-discover-abstract-apply cycle. Do not shortcut the method.
2. Capture all ideas: Document every idea with a brief rationale, even seemingly weak ones. Use structured format: idea statement, connection to the problem, potential testability. Quantity matters in this phase -- aim for 15-30 raw ideas per session.
3. Switch methods if needed: If ideas plateau after 15 minutes, switch to the backup method or select a new one from the Decision Framework. Plateaus are normal and expected -- they signal that one cognitive frame has been exhausted, not that the problem lacks solutions.
   • Transition criterion: Proceed to Phase 3 when you have at least 10-15 distinct ideas, or when two methods have been exhausted.

Phase 3: Connection Making (15-20% of session time)

1. Cluster related ideas: Group ideas by theme, mechanism, or approach. Name each cluster with a descriptive label. Typical sessions produce 3-6 clusters. Ideas that do not fit any cluster may represent novel directions worth exploring.
2. Identify bridges between clusters: Look for ideas that connect two or more clusters. These bridging ideas often represent the most innovative directions because they combine multiple insights from different cognitive frames.
3. Generate hybrid ideas: Deliberately combine elements from different clusters. Use Morphological Analysis on the clusters themselves: pick one element from cluster A, one from cluster B, and explore the combination. Often the most promising research direction is not any single idea but a synthesis of ideas from multiple clusters.
   • Transition criterion: Proceed to Phase 4 when ideas are organized into clusters with identified connections and at least 2-3 hybrid ideas generated.

Phase 4: Critical Evaluation (15-20% of session time)

1. Apply feasibility filters: For each promising idea or cluster, assess: Is it testable with available methods? Are the required resources accessible? What is the approximate timeline? What is the risk of failure? Rank ideas by feasibility and potential impact.
2. Use Six Thinking Hats for top candidates: For the 3-5 most promising directions, systematically evaluate from all six perspectives. White: what data exists? Red: what feels right or wrong about this? Black: what could go wrong? Yellow: what is the best outcome? Green: what variations exist? Blue: how should we proceed?
3. Identify contradictions and resolve with TRIZ: If promising ideas have inherent contradictions (improving one aspect worsens another), apply TRIZ inventive principles to find resolution strategies. Many apparently impossible ideas become feasible when the underlying contradiction is explicitly identified and resolved.
   • Transition criterion: Proceed to Phase 5 when top 3-5 ideas are evaluated and ranked by feasibility and potential impact.

Phase 5: Synthesis and Next Steps (10-15% of session time)

1. Prioritize directions: Select 2-3 research directions from the evaluated ideas. For each, articulate: the core idea, why it is promising, what would need to be true for it to work, and what the first experiment or investigation would be.
2. Define concrete follow-up actions: For each prioritized direction, specify: (a) literature search terms to check novelty and prior work, (b) preliminary feasibility experiments or pilot studies, (c) key assumptions that need validation before full commitment, and (d) timeline for initial investigation.
3. Document the session: Record the full brainstorming output: all ideas generated (including discarded ones), clusters identified, evaluation results, and prioritized directions with next steps. This document serves as a reference for future sessions and prevents re-exploration of already-considered ideas.
   • Completion criterion: Session is complete when 2-3 prioritized directions have defined next steps with owners and timelines.

Adaptive Strategies

When the session is not producing results, apply these interventions:

• If stuck with no ideas: Switch from analytical methods (SCAMPER, Morphological Analysis) to provocative methods (Provocation, Random Input). The block is usually caused by analytical thinking inhibiting creative thinking. A single absurd provocation often breaks the logjam.

• If ideas feel too safe or incremental: Increase the challenge level. Use Reverse Assumptions to flip the most fundamental assumption in the problem. Ask "what would a Nobel Prize-winning solution look like?" to raise the aspiration level. Apply TRIZ Ideal Final Result to envision the perfect outcome unconstrained by current limitations.

• If energy or engagement is dropping: Switch to a more interactive method. Random Input is particularly effective because each random stimulus creates a fresh starting point. Alternatively, take a brief break and return with Biomimicry -- exploring nature's solutions is inherently engaging.

• If generating many ideas but they feel disconnected: Pause divergent generation and spend time in Phase 3 (Connection Making). Map the ideas generated so far, look for clusters and bridges. Often the best idea is a combination of two mediocre ideas from different clusters.

• If one person is dominating the session: Impose structure with Six Thinking Hats, which forces rotation through perspectives. Alternatively, use silent brainstorming: each participant writes ideas independently for 5 minutes, then shares and builds on each other's written ideas.

Further Reading

• Lateral Thinking by Edward de Bono -- foundational work on deliberate creativity techniques including the Provocation (Po) method and Six Thinking Hats
• TRIZ: Theory of Inventive Problem Solving -- comprehensive resource for contradiction matrices, inventive principles, and ideal final result methodology
• Biomimicry Institute -- resources for nature-inspired design including the AskNature database of biological strategies
• Morphological Analysis (Fritz Zwicky) -- systematic exploration of multi-dimensional solution spaces, originally developed for astrophysics and engineering
• SCAMPER technique (Bob Eberle) -- structured checklist for creative modification of existing products, processes, or methods

Related Skills

• hypothesis-generation -- downstream from brainstorming; converts prioritized ideas into formal, testable hypotheses with experimental designs
• scientific-manuscript-writing -- downstream from ideation; structures brainstorming outcomes into research proposals and manuscripts
• peer-review-methodology -- provides critical evaluation frameworks that complement the Critical Evaluation phase of brainstorming