Skip to content

Editorial Standard — chipExercise session (chip-guided CBT exercise)

Purpose

A chipExercise session presents a CBT prompt -- a question, a situation classification, or a category-mapping task -- and a set of selectable chips from which the user chooses one or more answers. The session functions as guided self-assessment: it externalises a cognitive process (identifying triggers, classifying thoughts, mapping avoidance patterns) into a structured interaction that produces a data point for the programme. Unlike free-text journaling, chips reduce the cognitive load of self-disclosure and make the output consistent enough for downstream coaching logic to use.

chipExercise sessions appear across multiple CBT components: thought monitoring (identifying automatic thoughts by type), trigger mapping (which situations provoke distress), avoidance identification, and gratitude exercises. Each session covers one assessment dimension only.

Voice register

Default register is early habituation -- companion-forward, warm, specific. The intro paragraphs frame the exercise without over-explaining; they tell the user what they are about to do and why it matters for their programme. The knowledgeable guide is inviting the user into a concrete task, not asking them to perform vulnerability.

Chip labels themselves are register-neutral: they should be plain-language, recognisable, and non-clinical (not "hypervigilance" but "I keep checking whether the ringing has changed"). Avoid labelling options in a way that implies any selection is the right or wrong answer -- ACT design principle: multiple perspectives are valid.

For crisis-adjacent chipExercise sessions (e.g. spike situation mapping in early weeks): stay close to the habituation register but be more directive in the intro framing. The user is not there to feel supported -- they are there to produce a useful data point.

Evidence / IP

Grounded in docs/session-content-evidence-base.md. Key CBT components chipExercise sessions implement:

  • Thought monitoring / thought diary: rated 9/10 effectiveness by patients in the 2025 Tonndorf lecture (PMC12109689). The 25-component CBT review (vault note #586) identifies thought monitoring as a distinct component (not the same as thought challenging). chipExercise is the natural format for structured thought-classification exercises.
  • Trigger identification: part of psychoeducation components (vault note #586 category: Cognitive). Knowing which situations provoke distress is a prerequisite for graded exposure and avoidance-reduction work.
  • Avoidance mapping: from the 25-component review; also graded exposure component (Behavioural category). The chipExercise format operationalises the assessment stage before the experiment or exposure.
  • Gratitude exercises: from Tinnibot's MBCT module component. Three-item gratitude exercises have well-established evidence in positive psychology, and Tinnibot's integration into a tinnitus programme is documented.
  • Self-efficacy building: from Kalmeda Level 5 (evidence base section 3g). chipExercise sessions that ask users to identify what they have handled well serve the self-efficacy function.

IP: CBT techniques are academic and unprotectable. Chip content is authored by Naluma -- no reproduction of protected programme text required. The iCBT4Tinnitus 4C framework (Classify, Cope, Control, Communicate) is a structural device that can inform session design without reproducing its wording.

Length / reading level

  • Intro paragraphs (1-N): 2-4 sentences total across all intro paragraphs. State the task, say what it is for in the programme, and prompt the action. First intro sentence should be the task frame, not a warmth statement.
  • Chip labels: 2-8 words. Concise, plain-language, specific. Not abstract ("stress") -- specific ("noise that surprises me at work"). Users read chip labels at a glance.
  • Chip count: 4-8 chips is the working range. Below 4 limits usefulness of the assessment. Above 8 creates decision fatigue. For a max_selected of 3, 6-8 chips gives appropriate variation.
  • Reading level: Grade 8 or below. Chip labels should be even simpler -- Grade 6 if possible.

Editorial-required elements

  1. One assessment dimension per session. A chipExercise asking about sleep triggers AND social avoidance AND cognitive patterns covers three dimensions. Split it. Each session produces one output type -- the session's value field must represent a coherent, consistent data point for downstream logic.

  2. Chip value stability. The value field on each chip is the machine key persisted to the user-content row. Once a programme is live, changing chip values is a breaking change. Values should be set with care and treated as permanent: lowercase_snake_case, descriptive, not abbreviated.

  3. No implied correct answer. The intro framing and chip labels must not imply that selecting certain options is better or worse. "Which of these sounds familiar?" not "Which unhealthy patterns do you notice?" The user's self-knowledge is accurate; the exercise acknowledges it, not evaluates it.

  4. Intro must state the purpose. The user should understand what this data is used for, at least in general terms. "This helps us tailor the next few sessions to situations that matter for you" is sufficient. A chipExercise with no purpose statement asks the user to perform a task for no visible reason.

  5. Obeys ai-patterns-en.md. Intro copy and chip labels must obey all of ai-patterns-en.md. Chip labels are particularly at risk for generic wellness language ("feeling overwhelmed", "stress and anxiety") -- prefer tinnitus-specific, experiential language.

Examples

Good -- trigger-mapping chipExercise:

Intros: "Different situations make tinnitus harder to ignore. Select the ones that feel most relevant to you right now -- this shapes the next two sessions."

Chips: "Quiet rooms at night", "Trying to concentrate at work", "Background noise that drowns the ringing out", "Social events where I have to listen hard", "First thing in the morning", "Lying in bed"

Why this works: the intro is specific (shapes next two sessions), the chips are experiential and tinnitus-specific, each label is a recognisable lived moment rather than a clinical category, and no chip implies a wrong answer.

Bad -- thought-monitoring chipExercise:

❌ Intros: "As you continue your wellness journey, it's important to manage your tinnitus by identifying negative thought patterns that may be affecting your wellbeing."

❌ Chips: "Catastrophising", "Negative thinking", "Stress", "Anxiety", "Depression"

Why it fails: "wellness journey" and "manage your tinnitus" are banned under ai-patterns-en.md (Naluma-voice additions). The chip labels are clinical categories, not experiential descriptions -- a user who "catastrophises" does not think of it that way and cannot match themselves to the label. The intro frames all selections as "negative," which implies a correct answer (no selections) and undermines honest self-assessment.