@brainpilot/skills 0.0.6

package/skills/03_Cognitive_Psychology/signal-detection-analysis/SKILL.md

@@ -0,0 +1,300 @@
+---
+name: "signal-detection-analysis"
+description: "Domain-validated decision logic, formulas, and interpretation guidelines for applying Signal Detection Theory to cognitive science data"
+domain: "cognitive-psychology"
+version: "1.0.0"
+authors:
+ - "Claude (AI-assisted)"
+papers:
+ - "Green & Swets, 1966"
+ - "Macmillan & Creelman, 2005"
+ - "Stanislaw & Todorov, 1999"
+ - "Hautus, 1995"
+dependencies:
+ required:
+ - research-literacy
+review_status: "ai-generated"
+---
+
+# Signal Detection Analysis
+
+## Purpose
+
+This skill encodes expert methodological knowledge for applying Signal Detection Theory (SDT) to behavioral and cognitive science data. SDT separates an observer's perceptual sensitivity from their decision criterion -- a distinction that raw accuracy conflates. A competent programmer without cognitive science training would typically compute percent correct, missing the critical insight that two observers with identical accuracy can differ drastically in their ability to detect signals vs. their willingness to say "yes."
+
+## When to Use SDT (Not Simple Accuracy)
+
+Use SDT whenever:
+
+- Stimuli belong to two classes (signal vs. noise, old vs. new, present vs. absent) and the observer makes a binary classification
+- You need to distinguish **how well** someone can discriminate (sensitivity) from **how willing** they are to respond in a particular way (bias/criterion)
+- Response bias may differ across conditions, groups, or time points, making raw accuracy misleading
+- You want a measure that is independent of base rates and payoff structures
+
+Do **not** use standard SDT when:
+
+- There are more than two stimulus classes (use multi-class extensions or confusion matrices)
+- Responses are continuous rather than categorical (use regression-based approaches)
+- The task has no noise distribution (e.g., simple threshold detection with catch trials absent)
+
+## Research Planning Protocol
+
+Before executing the domain-specific steps below, you MUST:
+
+1. **State the research question** — What sensitivity or bias question is this SDT analysis addressing?
+2. **Justify the method choice** — Why SDT (not simple accuracy, logistic regression, etc.)? What alternatives were considered?
+3. **Declare expected outcomes** — Do you expect sensitivity differences, bias differences, or both?
+4. **Note assumptions and limitations** — What does SDT assume (e.g., Gaussian distributions, equal variance)? Where could it mislead?
+5. **Present the plan to the user and WAIT for confirmation** before proceeding.
+
+For detailed methodology guidance, see the `research-literacy` skill.
+
+
+## ⚠️ Verification Notice
+
+This skill was generated by AI from academic literature. All parameters, thresholds, and citations require independent verification before use in research. If you find errors, please [open an issue](https://github.com/HaoxuanLiTHUAI/awesome_cognitive_and_neuroscience_skills/issues).
+
+## Core Concepts
+
+### The 2x2 Response Matrix
+
+Every SDT analysis begins with classifying each trial into one of four categories:
+
+| | Signal Present | Signal Absent |
+|--|----------------|---------------|
+| **"Yes" Response** | Hit (H) | False Alarm (FA) |
+| **"No" Response** | Miss (M) | Correct Rejection (CR) |
+
+From these four cells, compute two rates:
+
+- **Hit Rate**: H / (H + M) = proportion of signal trials correctly identified
+- **False Alarm Rate**: FA / (FA + CR) = proportion of noise trials incorrectly called "signal"
+
+### Sensitivity: d' (d-prime)
+
+d' measures the distance between the signal and noise distributions in standard deviation units, assuming equal-variance Gaussian distributions (Green & Swets, 1966, Ch. 1):
+
+> **d' = z(Hit Rate) - z(False Alarm Rate)**
+
+where z() is the inverse of the standard normal CDF (the z-transform).
+
+**What d' values mean in practice** (Macmillan & Creelman, 2005; Table 1.1):
+
+| d' Value | Yes/No Interpretation | 2AFC % Correct | Practical Meaning |
+|----------|----------------------|----------------|-------------------|
+| 0 | Chance performance | 50% | No discrimination ability |
+| 0.5 | Low sensitivity | ~60% | Barely above chance |
+| 1.0 | Moderate sensitivity | ~69% | Often used as threshold (Green & Swets, 1966, Ch. 4) |
+| 2.0 | Good sensitivity | ~84% | Reliable discrimination |
+| 2.5 | High sensitivity | ~90% | Strong discrimination |
+| 3.0+ | Near-ceiling | >93% | Approaching perfect; check for floor/ceiling issues |
+
+The typical experimental range avoiding floor/ceiling effects is **d' = 0.5 to 2.5** (Macmillan & Creelman, 2005).
+
+### Bias Measures: c, beta, and c'
+
+SDT provides three interchangeable bias measures. The choice matters when d' varies across conditions.
+
+**Criterion location c** (Macmillan & Creelman, 2005, Ch. 2):
+
+> **c = -0.5 x [z(Hit Rate) + z(False Alarm Rate)]**
+
+- c = 0: unbiased (optimal for equal base rates and symmetric payoffs)
+- c > 0: conservative (tendency to say "no" / fewer false alarms, fewer hits)
+- c < 0: liberal (tendency to say "yes" / more hits, more false alarms)
+
+**Likelihood ratio beta** (Green & Swets, 1966, Ch. 1):
+
+> **ln(beta) = d' x c**
+
+- beta = 1: unbiased
+- beta > 1: conservative
+- beta < 1: liberal
+
+**Relative criterion c'** (Macmillan & Creelman, 2005, Ch. 2):
+
+> **c' = c / d'**
+
+Normalizes criterion placement by sensitivity; useful when comparing bias across conditions with different d' values.
+
+**Which bias measure to use** (Macmillan & Creelman, 2005, Ch. 2):
+
+- Use **c** as the default -- it is statistically independent of d', defined when d' = 0, and symmetric around chance
+- Use **beta** when testing whether observers approximate an optimal likelihood-ratio decision rule (e.g., recognition memory; Stretch & Wixted, 1998)
+- Use **c'** when you need to compare bias across conditions where d' changes substantially
+
+## Decision Logic: Choosing a Sensitivity Measure
+
+```
+Is the task a single-interval (yes/no) design?
+|
++-- YES --> Are assumptions of equal-variance Gaussian distributions met?
+| |
+| +-- YES --> Use d' = z(H) - z(FA) (Green & Swets, 1966)
+| |
+| +-- NO, distributions have unequal variance
+| | --> Use da with estimated variance ratio
+| | (Macmillan & Creelman, 2005, Ch. 3)
+| |
+| +-- NO, distributions are non-Gaussian or unknown
+| --> Use Az (area under the ROC curve)
+| (Swets, 1986; Macmillan & Creelman, 2005, Ch. 3)
+|
++-- NO --> Is it a two-interval forced choice (2AFC/2IFC)?
+ |
+ +-- YES --> d'(2AFC) = z(proportion correct) x sqrt(2)
+ | (Green & Swets, 1966, Ch. 6; Macmillan & Creelman, 2005, Ch. 5)
+ |
+ +-- NO --> Is it same-different or ABX?
+ |
+ +-- YES --> Use paradigm-specific formulas
+ | (see references/sdt-formulas.md)
+ |
+ +-- NO --> Is it a rating-scale (confidence) design?
+ |
+ +-- YES --> Construct ROC from rating data;
+ use Az or fit parametric model
+ (Macmillan & Creelman, 2005, Ch. 3)
+```
+
+### When to Use Az Instead of d'
+
+Use the area under the ROC curve (Az) when:
+
+1. You have **rating-scale data** (multiple confidence levels) and can construct a full ROC
+2. The **equal-variance assumption is violated** (common in recognition memory, where the zROC slope is typically ~0.80 rather than 1.0; Mickes, Wixted, & Wais, 2007; Ratcliff, Sheu, & Gronlund, 1992)
+3. You want a **distribution-free** sensitivity measure that does not assume Gaussian internals (Swets, 1986)
+
+**AUC benchmarks** (Swets, Dawes, & Monahan, 2000):
+
+| AUC Range | Interpretation |
+|-----------|----------------|
+| 0.50 | Chance (no discrimination) |
+| 0.70 - 0.80 | Fair diagnostic accuracy |
+| 0.80 - 0.90 | Good diagnostic accuracy |
+| 0.90 - 1.00 | Excellent diagnostic accuracy |
+
+## Common Paradigms
+
+### Yes/No Detection
+
+The canonical SDT paradigm. On each trial, either a signal or noise is presented; the observer responds "yes" (signal present) or "no" (signal absent). Yields H and FA rates directly.
+
+- d' = z(H) - z(FA)
+- Bias c = -0.5 x [z(H) + z(FA)]
+
+### Two-Alternative Forced Choice (2AFC / 2IFC)
+
+Two intervals are presented (one signal, one noise); the observer selects the signal interval. Only proportion correct is measured; there is no independent FA rate, and no bias measure can be computed.
+
+- d'(2AFC) = z(proportion correct) x sqrt(2) (Green & Swets, 1966, Ch. 6)
+- d'(2AFC) = d'(yes/no) x sqrt(2) (Macmillan & Creelman, 2005, Ch. 5)
+
+**Critical domain pitfall**: A task where the observer chooses between two labels (e.g., "left" or "right") on a single stimulus is **not** a 2AFC -- it is a yes/no task in disguise (Macmillan & Creelman, 2005). True 2AFC requires two temporal or spatial intervals.
+
+### Rating Scale (Confidence Ratings)
+
+Observers make a detection judgment plus a confidence rating (e.g., 1-6 scale from "sure noise" to "sure signal"). Each confidence boundary yields a separate (H, FA) pair, constructing a multi-point ROC.
+
+- Fit with parametric (Gaussian) or nonparametric methods
+- Compute Az from the fitted ROC
+- The zROC slope estimates the variance ratio of the two distributions
+
+### Same-Different
+
+Two stimuli are presented; the observer judges "same" or "different." Two observer models exist (Macmillan & Creelman, 2005, Ch. 6):
+
+- **Independent observations model**: observer compares each stimulus to an internal criterion
+- **Differencing model**: observer computes the difference between the two percepts
+
+These yield different d' formulas; see `references/sdt-formulas.md`.
+
+### ABX (Oddity / AX)
+
+Stimulus A, then B, then X (which matches A or B); the observer identifies X. Sensitivity depends on assumed observer strategy (Macmillan & Creelman, 2005, Ch. 6). See `references/sdt-formulas.md`.
+
+## Handling Extreme Hit/False Alarm Rates
+
+When H = 1.0 or FA = 0.0, z-scores become infinite and d' is undefined. This is a **common computational pitfall** that requires correction.
+
+### Correction Methods
+
+**1. The 1/(2N) rule** (Macmillan & Kaplan, 1985):
+
+- Replace 0 with 0.5/N
+- Replace 1 with (N - 0.5)/N
+- Where N = number of signal trials (for H) or noise trials (for FA)
+- Applied **only** to extreme values
+
+**2. The log-linear rule** (Hautus, 1995) -- **recommended**:
+
+- Add 0.5 to every cell in the 2x2 matrix (hits, misses, FA, CR) **before** computing rates
+- Applied to **all** cells, regardless of whether extremes are present
+- Produces less biased estimates that consistently underestimate true d' (Hautus, 1995)
+
+**Which to use**: The log-linear rule is preferred because it produces less biased d' estimates and avoids the asymmetric bias of the 1/(2N) rule, which can either over- or underestimate d' (Hautus, 1995). Apply the log-linear correction **routinely**, not just when extremes occur, for consistency across participants and conditions.
+
+## The Unequal-Variance Problem
+
+### Why It Matters
+
+Standard d' assumes signal and noise distributions have equal variance. In recognition memory, this assumption is **routinely violated**: zROC slopes are typically **~0.80** (not 1.0), indicating the old-item (target) distribution has ~25% more variance than the new-item (lure) distribution (Ratcliff, Sheu, & Gronlund, 1992; Mickes, Wixted, & Wais, 2007).
+
+### Consequences of Ignoring It
+
+If variances are unequal and you compute standard d', the measure is **not criterion-free** -- it will vary with criterion placement even if true sensitivity is constant (Macmillan & Creelman, 2005, Ch. 3).
+
+### What to Do
+
+1. Collect rating-scale data to construct a zROC
+2. If the zROC slope deviates from 1.0, use the unequal-variance model
+3. Compute da (the unequal-variance sensitivity measure); see `references/sdt-formulas.md`
+
+## Common Pitfalls
+
+1. **Using percent correct instead of d'**: Percent correct confounds sensitivity and bias. Two observers with identical discrimination ability but different criteria will have different accuracy scores (Green & Swets, 1966, Ch. 1).
+
+2. **Treating a single-stimulus forced choice as 2AFC**: If only one stimulus is presented per trial and the observer picks a label, this is a yes/no design, not 2AFC. Using the 2AFC formula will yield incorrect d' values (Macmillan & Creelman, 2005).
+
+3. **Ignoring extreme rate corrections**: Computing d' without correcting H = 1 or FA = 0 produces infinite values. Always apply the log-linear correction (Hautus, 1995).
+
+4. **Assuming equal variance in recognition memory**: Recognition memory data almost always show unequal variance (zROC slope ~0.80). Standard d' is not criterion-free in this domain (Ratcliff, Sheu, & Gronlund, 1992).
+
+5. **Interpreting c as "response bias" without checking**: c measures where the criterion is placed relative to distributions, not why it is placed there. A shift in c can reflect rational adaptation to base rates, not irrational bias (Macmillan & Creelman, 2005, Ch. 2).
+
+6. **Comparing d' across paradigms without conversion**: d' values from yes/no and 2AFC designs are not directly comparable. d'(2AFC) = d'(yes/no) x sqrt(2). Failure to convert leads to erroneous sensitivity comparisons (Green & Swets, 1966, Ch. 6).
+
+7. **Averaging d' across participants without caution**: d' is nonlinearly related to H and FA rates. Averaging H and FA rates first, then computing d', gives different results than averaging individual d' values. The appropriate method depends on the research question (Macmillan & Creelman, 2005, Ch. 8).
+
+## Minimum Reporting Checklist
+
+Based on Macmillan & Creelman (2005) and Stanislaw & Todorov (1999):
+
+- [ ] Paradigm type (yes/no, 2AFC, rating, same-different, ABX)
+- [ ] Number of signal and noise trials per condition
+- [ ] Hit rate and false alarm rate (or full rating distribution)
+- [ ] Correction method used for extreme proportions (log-linear or 1/2N)
+- [ ] Sensitivity measure (d', da, Az) with justification for choice
+- [ ] Bias measure (c, beta, c') with justification for choice
+- [ ] Whether equal- or unequal-variance model was used (and estimated variance ratio if unequal)
+- [ ] If rating data: ROC and/or zROC plot with slope reported
+- [ ] Statistical tests on SDT measures (not on raw accuracy)
+- [ ] Software and version used for computation
+
+## References
+
+- Green, D. M., & Swets, J. A. (1966). *Signal detection theory and psychophysics*. New York: Wiley.
+- Hautus, M. J. (1995). Corrections for extreme proportions and their biasing effects on estimated values of d'. *Behavior Research Methods, Instruments, & Computers*, 27, 46-51.
+- Macmillan, N. A., & Creelman, C. D. (2005). *Detection theory: A user's guide* (2nd ed.). Mahwah, NJ: Erlbaum.
+- Macmillan, N. A., & Kaplan, H. L. (1985). Detection theory analysis of group data. *Psychological Bulletin*, 98, 185-199.
+- Maniscalco, B., & Lau, H. (2012). A signal detection theoretic approach for estimating metacognitive sensitivity from confidence ratings. *Consciousness and Cognition*, 21, 422-430.
+- Mickes, L., Wixted, J. T., & Wais, P. E. (2007). A direct test of the unequal-variance signal detection model of recognition memory. *Psychonomic Bulletin & Review*, 14, 858-865.
+- Ratcliff, R., Sheu, C. F., & Gronlund, S. D. (1992). Testing global memory models using ROC curves. *Psychological Review*, 99, 518-535.
+- Stanislaw, H., & Todorov, N. (1999). Calculation of signal detection theory measures. *Behavior Research Methods, Instruments, & Computers*, 31, 137-149.
+- Stretch, V., & Wixted, J. T. (1998). On the difference between strength-based and frequency-based mirror effects in recognition memory. *Journal of Experimental Psychology: Learning, Memory, and Cognition*, 24, 1379-1396.
+- Swets, J. A. (1986). Indices of discrimination or diagnostic accuracy. *Psychological Bulletin*, 99, 100-117.
+- Swets, J. A. (1988). Measuring the accuracy of diagnostic systems. *Science*, 240, 1285-1293.
+- Swets, J. A., Dawes, R. M., & Monahan, J. (2000). Psychological science can improve diagnostic decisions. *Psychological Science in the Public Interest*, 1, 1-26.
+
+See `references/sdt-formulas.md` for detailed mathematical formulas and lookup tables.
+See `references/application-guide.md` for domain-specific applications.

package/skills/03_Cognitive_Psychology/signal-detection-analysis/references/application-guide.md

@@ -0,0 +1,278 @@
+# SDT Application Guide: Domain-Specific Uses
+
+This reference file supplements the main `SKILL.md` with detailed guidance on applying Signal Detection Theory in specific cognitive science domains.
+
+## 1. Recognition Memory
+
+### The Old/New Paradigm
+
+The most common application of SDT in memory research. Participants study a list of items, then are tested with a mix of old (studied) and new (unstudied) items, responding "old" or "new."
+
+**Mapping to SDT** (Macmillan & Creelman, 2005, Ch. 2):
+- Signal = old item; Noise = new item
+- Hit = correctly responding "old" to an old item
+- False Alarm = incorrectly responding "old" to a new item
+- d' reflects **memory strength** (discriminability of old from new)
+- c reflects **response criterion** (willingness to endorse items as "old")
+
+### The Mirror Effect
+
+A robust finding: variables that increase hit rates also decrease false alarm rates (e.g., high-frequency words produce both lower H and higher FA than low-frequency words). This "mirror effect" (Glanzer & Adams, 1985) is naturally explained by SDT if both distributions shift (Stretch & Wixted, 1998).
+
+### Unequal Variance: The Default in Recognition Memory
+
+Standard d' assumes equal variance, but recognition memory data **almost always** violate this assumption (Ratcliff, Sheu, & Gronlund, 1992):
+
+- **zROC slopes** in recognition memory are typically **0.80** (range: 0.70-0.90 across studies; Ratcliff, Sheu, & Gronlund, 1992; Mickes, Wixted, & Wais, 2007)
+- This means sigma_old / sigma_new ~ **1.25** (old items have more variable memory strength)
+- The encoding variability hypothesis explains this: items are encoded with variable strength, increasing the variance of the old-item distribution (Wixted, 2007)
+
+**Practical recommendation**: When analyzing recognition memory data:
+1. Collect confidence ratings (e.g., 6-point scale: sure new, probably new, maybe new, maybe old, probably old, sure old)
+2. Construct the zROC and estimate the slope
+3. If slope deviates from 1.0 (it will), compute da rather than d'
+4. Report the slope and variance ratio alongside sensitivity
+
+### The Remember/Know Paradigm
+
+Participants indicate whether they "remember" specific details (recollection) or "know" the item was studied based on familiarity. Under the dual-process SDT model (Yonelinas, 1994):
+
+- **Recollection** (R) is modeled as a threshold process: P(R) = proportion of old items exceeding a recollection threshold
+- **Familiarity** is modeled as a continuous Gaussian SDT process: d'(familiarity) is estimated after removing recollected items
+
+**Estimation** (Yonelinas, 2002):
+- Recollection = P("remember" | old) - P("remember" | new)
+- Familiarity d' = z(P("know" | old) / (1 - P("remember" | old))) - z(P("know" | new) / (1 - P("remember" | new)))
+
+**Critical pitfall**: The independence assumption underlying these formulas (that "know" responses come only from non-recollected items) is debated. The redundancy model (Wixted & Mickes, 2010) argues recollection and familiarity are not independent. Report which model was assumed.
+
+## 2. Perception and Psychophysics
+
+### Threshold Estimation
+
+SDT was originally developed for perceptual detection (Green & Swets, 1966). The "threshold" in SDT is typically defined as the stimulus intensity yielding a criterion level of d':
+
+- **d' = 1.0** is the most common threshold definition in yes/no tasks (Green & Swets, 1966, Ch. 4)
+- **76% correct** is the equivalent threshold in 2AFC (corresponding to d'(2AFC) ~ 1.0; Green & Swets, 1966, Ch. 6)
+
+Other threshold conventions exist:
+- 75% correct in 2AFC (d' ~ 0.95) -- used in some adaptive staircase methods
+- d' = 2.0 for clinical hearing tests -- a more conservative threshold
+
+### Psychometric Functions and SDT
+
+The psychometric function (proportion correct vs. stimulus intensity) is related to SDT:
+
+- At each stimulus intensity, the observer has a particular d'
+- Plotting d' vs. stimulus intensity yields the psychometric function in SDT units
+- The slope of this function reflects how rapidly sensitivity changes with stimulus strength
+
+**Adaptive methods** (staircases, QUEST, PEST) estimate the stimulus level corresponding to a target d' or proportion correct (Watson & Pelli, 1983; Taylor & Creelman, 1967).
+
+### Signal Detection in Vision Research
+
+Common visual paradigms and their SDT implementations:
+
+| Task | SDT Design | Sensitivity Measure |
+|------|-----------|-------------------|
+| Contrast detection | Yes/no with noise intervals | d' |
+| Orientation discrimination | 2AFC (which interval had tilted grating?) | d'(2AFC) |
+| Change detection | Yes/no ("did something change?") | d' or Cowan's K (Cowan, 2001) |
+| Visual search | Yes/no ("target present?") | d' (separate for each set size) |
+
+**Visual search and SDT**: In visual search, d' typically decreases linearly with the logarithm of set size for inefficient search, but remains constant for efficient ("pop-out") search (Palmer, Verghese, & Pavel, 2000).
+
+### Signal Detection in Auditory Research
+
+Common auditory paradigms:
+
+| Task | SDT Design | Notes |
+|------|-----------|-------|
+| Tone detection in noise | Yes/no or 2IFC | Classic SDT paradigm (Green & Swets, 1966) |
+| Frequency discrimination | 2IFC or same-different | Use appropriate paradigm-specific formula |
+| Speech in noise | Open-set (word identification) | SDT less applicable; use percent correct or SRT |
+| Audiometric threshold | Adaptive staircase | Target d' varies by clinical standard |
+
+## 3. Clinical Psychology and Diagnostic Accuracy
+
+### SDT Framework for Diagnosis
+
+Any diagnostic decision can be framed as signal detection (Swets, 1988; Swets, Dawes, & Monahan, 2000):
+
+- **Signal** = condition present (e.g., disease, disorder, threat)
+- **Noise** = condition absent
+- **Hit** = true positive (correctly detecting condition)
+- **False Alarm** = false positive (incorrectly diagnosing condition when absent)
+- **Sensitivity** (d' or AUC) = diagnostic accuracy of the test/clinician
+- **Criterion** (c) = diagnostic threshold (how much evidence is required before diagnosing)
+
+### AUC Benchmarks for Clinical Instruments
+
+From Swets (1988) and Swets, Dawes, & Monahan (2000):
+
+| AUC | Clinical Interpretation |
+|-----|------------------------|
+| 0.50 | No diagnostic value (chance) |
+| 0.50 - 0.70 | Low accuracy; rarely useful clinically |
+| 0.70 - 0.80 | Fair accuracy; may be useful as one of several indicators |
+| 0.80 - 0.90 | Good accuracy; clinically useful |
+| 0.90 - 1.00 | Excellent accuracy; high diagnostic value |
+
+**Benchmark context**: Strong psychological assessment measures routinely yield AUC estimates around 0.70 (Youngstrom, 2014). An AUC of 0.80 is a reasonable target for clinically useful individual-level prediction.
+
+### Clinical Applications
+
+**Violence risk assessment** (Swets, Dawes, & Monahan, 2000):
+- Actuarial risk instruments (e.g., VRAG) achieve AUC ~ 0.70-0.76
+- Clinical judgment alone: AUC ~ 0.55-0.65
+- SDT analysis reveals that clinicians tend to have a conservative criterion (low FA but many misses)
+
+**Depression screening** (e.g., PHQ-9):
+- At cutoff score >= 10: sensitivity ~ 0.88, specificity ~ 0.88 (Kroenke, Spitzer, & Williams, 2001)
+- The cutoff score represents the criterion in SDT terms
+- Different cutoffs trade off sensitivity vs. specificity along the ROC curve
+
+**Medical imaging** (Swets, 1988):
+- Radiologist mammography reading: AUC ~ 0.80-0.90
+- Computer-aided detection systems: AUC ~ 0.85-0.95
+- SDT separates the radiologist's perceptual ability from their decision threshold
+
+### Base Rate Effects on Clinical Decisions
+
+A critical SDT insight for clinical work: the **optimal criterion shifts with base rate** (Green & Swets, 1966, Ch. 2).
+
+- In rare conditions (low base rate), the optimal criterion is conservative (high c)
+- In common conditions (high base rate), the optimal criterion is liberal (low c)
+- Clinicians often fail to adjust sufficiently for base rates, leading to systematic over-diagnosis of rare conditions and under-diagnosis of common ones (Swets, Dawes, & Monahan, 2000)
+
+**Positive predictive value** (PPV) depends on base rate even when sensitivity and specificity are constant:
+
+```
+PPV = (sensitivity * base_rate) / (sensitivity * base_rate + (1 - specificity) * (1 - base_rate))
+```
+
+For a test with sensitivity = 0.90 and specificity = 0.90:
+- Base rate 50%: PPV = 0.90
+- Base rate 10%: PPV = 0.50
+- Base rate 1%: PPV = 0.08
+
+This dramatic drop in PPV at low base rates is a key SDT-informed insight that clinicians without SDT training often miss.
+
+## 4. Metacognition and Confidence Ratings
+
+### Type 1 vs. Type 2 SDT
+
+**Type 1 SDT**: How well does the observer discriminate external stimulus states (signal vs. noise)?
+- Measures: d', c, Az
+
+**Type 2 SDT**: How well does the observer discriminate their own correct from incorrect responses?
+- Measures: type 2 d', type 2 AUC, meta-d' (Maniscalco & Lau, 2012)
+
+### Meta-d' (Maniscalco & Lau, 2012)
+
+Meta-d' is the type 1 d' that would produce the observed type 2 (metacognitive) performance under ideal SDT assumptions. It answers: "How much signal-to-noise information is available for metacognitive judgments?"
+
+**Key properties**:
+- Meta-d' is expressed in the same units as d', enabling direct comparison
+- If meta-d' = d': metacognition is ideal (all information available for the type 1 task is also used for metacognition)
+- If meta-d' < d': metacognitive inefficiency (some information is lost or unavailable for confidence judgments)
+- If meta-d' > d' (rare): suggests additional information sources for metacognition beyond the type 1 decision variable
+
+**Metacognitive efficiency** (M-ratio):
+
+```
+M-ratio = meta-d' / d'
+```
+
+- M-ratio = 1: ideal metacognition
+- M-ratio < 1: suboptimal metacognition (typical finding; Maniscalco & Lau, 2012)
+- Typical values in perceptual tasks: **M-ratio ~ 0.6 - 0.8** (Fleming & Lau, 2014)
+
+### Computing Meta-d'
+
+Meta-d' cannot be computed with a simple formula. It requires fitting a model to the type 2 data (confidence rating distributions conditioned on accuracy). Two approaches:
+
+1. **Maximum likelihood estimation** (Maniscalco & Lau, 2012): fit meta-d' to maximize the likelihood of observed type 2 hit and FA rates at each confidence level
+2. **Hierarchical Bayesian estimation** (Fleming, 2017): preferred for group-level analysis; handles individual differences and small trial counts better
+
+Software: Available at https://github.com/smfleming/HMeta-d (MATLAB and Python).
+
+### Confidence Calibration vs. Metacognitive Sensitivity
+
+These are distinct concepts often confused:
+
+- **Calibration**: Do confidence ratings match objective accuracy? (e.g., when someone says "80% sure," are they correct 80% of the time?)
+- **Resolution/sensitivity**: Can the observer distinguish correct from incorrect responses, regardless of absolute calibration?
+
+SDT-based measures (meta-d', type 2 AUC) measure **resolution**, not calibration. Calibration requires separate analysis (e.g., calibration curves).
+
+## 5. Eyewitness Identification
+
+### SDT in Lineup Research
+
+SDT has been applied to eyewitness identification (Wixted & Mickes, 2012):
+
+- **Signal** = target-present lineup
+- **Noise** = target-absent lineup
+- **Hit** = correct identification from target-present lineup
+- **False Alarm** = incorrect identification from target-absent lineup (filler or foil identification)
+
+### Lineup Procedure Effects
+
+SDT analysis reveals that different lineup procedures (simultaneous vs. sequential) primarily affect criterion placement, not discriminability (Wixted & Mickes, 2012):
+
+- **Simultaneous lineups**: more liberal criterion (higher H and higher FA)
+- **Sequential lineups**: more conservative criterion (lower H and lower FA)
+- **d' (discriminability)**: similar across procedures in many studies
+
+This SDT insight has important policy implications: sequential lineups were once recommended as superior, but SDT analysis shows they primarily make witnesses more conservative, not more accurate at discrimination (Wixted & Mickes, 2012).
+
+## 6. Decision Making Under Uncertainty
+
+### SDT and Optimal Decision Rules
+
+The optimal SDT observer sets the criterion to maximize expected utility (Green & Swets, 1966, Ch. 2):
+
+```
+beta_optimal = [P(noise) / P(signal)] * [(V_CR + C_FA) / (V_H + C_M)]
+```
+
+Where:
+- P(noise), P(signal) = prior probabilities
+- V_CR, V_H = values of correct decisions
+- C_FA, C_M = costs of incorrect decisions
+
+When base rates are equal and payoffs symmetric, beta_optimal = 1 (unbiased).
+
+### SDT and the Utility of Caution
+
+A key insight from "utilized" SDT (Lynn & Barrett, 2014): maximizing accuracy and maximizing utility are **not** always the same. There are situations where the optimal strategy is deliberately inaccurate:
+
+- When misses are far more costly than false alarms (e.g., smoke detector design), the optimal criterion is very liberal, maximizing hits at the expense of many false alarms
+- When false alarms are far more costly than misses (e.g., criminal conviction), the optimal criterion is very conservative
+
+## References
+
+- Cowan, N. (2001). The magical number 4 in short-term memory. *Behavioral and Brain Sciences*, 24, 87-114.
+- Fleming, S. M. (2017). HMeta-d: Hierarchical Bayesian estimation of metacognitive efficiency from confidence ratings. *Neuroscience of Consciousness*, 3(1), nix007.
+- Fleming, S. M., & Lau, H. C. (2014). How to measure metacognition. *Frontiers in Human Neuroscience*, 8, 443.
+- Glanzer, M., & Adams, J. K. (1985). The mirror effect in recognition memory. *Memory & Cognition*, 13, 8-20.
+- Green, D. M., & Swets, J. A. (1966). *Signal detection theory and psychophysics*. New York: Wiley.
+- Kroenke, K., Spitzer, R. L., & Williams, J. B. (2001). The PHQ-9: Validity of a brief depression severity measure. *Journal of General Internal Medicine*, 16, 606-613.
+- Lynn, S. K., & Barrett, L. F. (2014). "Utilizing" signal detection theory. *Psychological Science*, 25, 1663-1673.
+- Macmillan, N. A., & Creelman, C. D. (2005). *Detection theory: A user's guide* (2nd ed.). Mahwah, NJ: Erlbaum.
+- Maniscalco, B., & Lau, H. (2012). A signal detection theoretic approach for estimating metacognitive sensitivity from confidence ratings. *Consciousness and Cognition*, 21, 422-430.
+- Mickes, L., Wixted, J. T., & Wais, P. E. (2007). A direct test of the unequal-variance signal detection model of recognition memory. *Psychonomic Bulletin & Review*, 14, 858-865.
+- Palmer, J., Verghese, P., & Pavel, M. (2000). The psychophysics of visual search. *Vision Research*, 40, 1227-1268.
+- Ratcliff, R., Sheu, C. F., & Gronlund, S. D. (1992). Testing global memory models using ROC curves. *Psychological Review*, 99, 518-535.
+- Stretch, V., & Wixted, J. T. (1998). On the difference between strength-based and frequency-based mirror effects in recognition memory. *Journal of Experimental Psychology: Learning, Memory, and Cognition*, 24, 1379-1396.
+- Swets, J. A. (1988). Measuring the accuracy of diagnostic systems. *Science*, 240, 1285-1293.
+- Swets, J. A., Dawes, R. M., & Monahan, J. (2000). Psychological science can improve diagnostic decisions. *Psychological Science in the Public Interest*, 1, 1-26.
+- Taylor, M. M., & Creelman, C. D. (1967). PEST: Efficient estimates on probability functions. *Journal of the Acoustical Society of America*, 41, 782-787.
+- Watson, A. B., & Pelli, D. G. (1983). QUEST: A Bayesian adaptive psychometric method. *Perception & Psychophysics*, 33, 113-120.
+- Wixted, J. T. (2007). Dual-process theory and signal-detection theory of recognition memory. *Psychological Review*, 114, 152-176.
+- Wixted, J. T., & Mickes, L. (2010). A continuous dual-process model of remember/know judgments. *Psychological Review*, 117, 1025-1054.
+- Wixted, J. T., & Mickes, L. (2012). The field of eyewitness memory should abandon probative value and embrace receiver operating characteristic analysis. *Perspectives on Psychological Science*, 7, 275-278.
+- Yonelinas, A. P. (1994). Receiver-operating characteristics in recognition memory: Evidence for a dual-process model. *Journal of Experimental Psychology: Learning, Memory, and Cognition*, 20, 1341-1354.
+- Yonelinas, A. P. (2002). The nature of recollection and familiarity: A review of 30 years of research. *Journal of Memory and Language*, 46, 441-517.
+- Youngstrom, E. A. (2014). A primer on receiver operating characteristic analysis and diagnostic efficiency statistics for pediatric psychology. *Journal of Pediatric Psychology*, 39, 204-221.