Skip to content
Buy one Spikeling unit units

Neural-data analysis

Curriculum status

These entries are architecture specifications, not final lesson plans. They define teaching intent, logistics, outputs and scientific boundaries; release-ready settings, worksheets, notebooks and answer keys still require empirical validation.

Return to the module catalogue

Family source cluster: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]

DAT-01 — Import, data structure and metadata

Concept / theme A recording is usable only when columns, units, timebase, conditions and provenance are explicit.
Audience / context Undergraduate. Formal practical, workshop or modular course. Prior modules: FND-04, EPH-04.
Logistics 45 min; 1–2 per computer; boards: 0–1; software: Jupyter; GUI for acquisition/QC; prepared dataset: Yes: generated earlier or prepared. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Jupyter analysis. Stages: 6–12.
Spikeling relationship 5 — Recorded-dataset analysis; 3 — Hybrid board + Jupyter

Learning outcomes.

  • Identify and predict the principal behaviour described in a recording is usable only when columns, units, timebase, conditions and provenance are explicit.

  • Configure or document load csv, inspect schema and construct a metadata dictionary and record columns, dtypes, missing values, dt and condition labels.

  • Measure, calculate or compare load csv, inspect schema and construct a metadata dictionary using an explicit operational rule.

  • Interpret the result and state why do not infer missing units or sample interval from column names alone

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Load CSV, inspect schema and construct a metadata dictionary.

4. Acquire or inspect columns, dtypes, missing values, dt and condition labels.

5. Produce validated data object and metadata record.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Do not infer missing units or sample interval from column names alone.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate load csv, inspect schema and construct a metadata dictionary. The reusable output is validated data object and metadata record.

Sources / connections / priority. Sources: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]. Natural follow-ons: DAT-02, MET-06. Development priority: Core module. Boundary: Do not infer missing units or sample interval from column names alone.

DAT-02 — Plotting and quality control

Concept / theme Raw visualisation should precede feature extraction and statistical testing.
Audience / context Undergraduate. Formal practical, workshop or modular course. Prior modules: DAT-01.
Logistics 45 min; 1–2 per computer; boards: 0–1; software: Jupyter; GUI for acquisition/QC; prepared dataset: Yes: generated earlier or prepared. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Jupyter analysis. Stages: 6–12.
Spikeling relationship 5 — Recorded-dataset analysis; 3 — Hybrid board + Jupyter

Learning outcomes.

  • Identify and predict the principal behaviour described in raw visualisation should precede feature extraction and statistical testing.

  • Configure or document plot aligned channels, inspect baseline, clipping, discontinuities and trigger consistency and record Vm, current, stimulus, synaptic channels, trigger.

  • Measure, calculate or compare plot aligned channels, inspect baseline, clipping, discontinuities and trigger consistency using an explicit operational rule.

  • Interpret the result and state why a visually plausible trace can still have timing or metadata errors

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Plot aligned channels, inspect baseline, clipping, discontinuities and trigger consistency.

4. Acquire or inspect Vm, current, stimulus, synaptic channels, trigger.

5. Produce qc figure and accept/reject/flag log.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: A visually plausible trace can still have timing or metadata errors.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate plot aligned channels, inspect baseline, clipping, discontinuities and trigger consistency. The reusable output is qc figure and accept/reject/flag log.

Sources / connections / priority. Sources: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]. Natural follow-ons: DAT-03, EPH-06, MET-04. Development priority: Core module. Boundary: A visually plausible trace can still have timing or metadata errors.

DAT-03 — Spike detection and validation

Concept / theme Detection is an operational algorithm whose errors depend on threshold, noise and refractory settings.
Audience / context Undergraduate. Formal practical, workshop or modular course. Prior modules: DAT-02, EPH-05.
Logistics 60 min; 1–2 per computer; boards: 0–1; software: Jupyter; GUI for acquisition/QC; prepared dataset: Yes: generated earlier or prepared. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Jupyter analysis. Stages: 6–12.
Spikeling relationship 5 — Recorded-dataset analysis; 3 — Hybrid board + Jupyter; 6 — GUI simulation workflow

Learning outcomes.

  • Identify and predict the principal behaviour described in detection is an operational algorithm whose errors depend on threshold, noise and refractory settings.

  • Configure or document implement upward-crossing or peak detection and compare against known/synthetic events and record detected times, threshold, false positives/negatives.

  • Measure, calculate or compare implement upward-crossing or peak detection and compare against known/synthetic events using an explicit operational rule.

  • Interpret the result and state why detected events are algorithm outputs; retain uncertainty and ground-truth checks where available

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Implement upward-crossing or peak detection and compare against known/synthetic events.

4. Acquire or inspect detected times, threshold, false positives/negatives.

5. Produce detection performance table and selected parameters.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Detected events are algorithm outputs; retain uncertainty and ground-truth checks where available.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate implement upward-crossing or peak detection and compare against known/synthetic events. The reusable output is detection performance table and selected parameters.

Sources / connections / priority. Sources: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]. Natural follow-ons: DAT-04, IMG-05, EXT-04. Development priority: Core module. Boundary: Detected events are algorithm outputs; retain uncertainty and ground-truth checks where available.

DAT-04 — Firing rate and interspike intervals

Concept / theme Spike trains can be summarised by counts, rates and interval distributions at different timescales.
Audience / context Undergraduate. Formal practical, workshop or modular course. Prior modules: DAT-03, EPH-04.
Logistics 60 min; 1–2 per computer; boards: 0–1; software: Jupyter; GUI for acquisition/QC; prepared dataset: Yes: generated earlier or prepared. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Jupyter analysis. Stages: 6–12.
Spikeling relationship 5 — Recorded-dataset analysis; 3 — Hybrid board + Jupyter

Learning outcomes.

  • Identify and predict the principal behaviour described in spike trains can be summarised by counts, rates and interval distributions at different timescales.

  • Configure or document compute trial rate, instantaneous/filtered rate and isis and record spike times, window/kernel, rate and isi.

  • Measure, calculate or compare compute trial rate, instantaneous/filtered rate and isis using an explicit operational rule.

  • Interpret the result and state why rates depend on estimator/window; validated time units are mandatory

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Compute trial rate, instantaneous/filtered rate and ISIs.

4. Acquire or inspect spike times, window/kernel, rate and isi.

5. Produce rate panels and isi distribution.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Rates depend on estimator/window; validated time units are mandatory.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate compute trial rate, instantaneous/filtered rate and isis. The reusable output is rate panels and isi distribution.

Sources / connections / priority. Sources: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]. Natural follow-ons: DAT-05, STA-01. Development priority: Core module. Boundary: Rates depend on estimator/window; validated time units are mandatory.

DAT-05 — Latency, adaptation and burst features

Concept / theme Feature extraction turns qualitative patterns into reproducible operational measurements.
Audience / context Advanced undergraduate. Formal practical, workshop or modular course. Prior modules: DAT-03, DAT-04.
Logistics 60 min; 1–2 per computer; boards: 0–1; software: Jupyter; GUI for acquisition/QC; prepared dataset: Yes: generated earlier or prepared. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Jupyter analysis. Stages: 6–12.
Spikeling relationship 5 — Recorded-dataset analysis; 3 — Hybrid board + Jupyter

Learning outcomes.

  • Identify and predict the principal behaviour described in feature extraction turns qualitative patterns into reproducible operational measurements.

  • Configure or document calculate first-spike latency, adaptation indices and burst criteria and record spike times, stimulus onset, burst thresholds.

  • Measure, calculate or compare calculate first-spike latency, adaptation indices and burst criteria using an explicit operational rule.

  • Interpret the result and state why feature thresholds must be justified and sensitivity-checked

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Calculate first-spike latency, adaptation indices and burst criteria.

4. Acquire or inspect spike times, stimulus onset, burst thresholds.

5. Produce feature table with definitions.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Feature thresholds must be justified and sensitivity-checked.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate calculate first-spike latency, adaptation indices and burst criteria. The reusable output is feature table with definitions.

Sources / connections / priority. Sources: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]. Natural follow-ons: CMP-03, STA-03. Development priority: High-priority extension. Boundary: Feature thresholds must be justified and sensitivity-checked.

DAT-06 — Trial alignment, rasters and PSTHs

Concept / theme Repeated trials reveal response timing, reliability and condition-dependent structure.
Audience / context Advanced undergraduate. Formal practical, workshop or modular course. Prior modules: DAT-01, DAT-03, EPH-04.
Logistics 60 min; 1–2 per computer; boards: 0–1; software: Jupyter; GUI for acquisition/QC; prepared dataset: Yes: generated earlier or prepared. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Jupyter analysis. Stages: 6–12.
Spikeling relationship 5 — Recorded-dataset analysis; 3 — Hybrid board + Jupyter

Learning outcomes.

  • Identify and predict the principal behaviour described in repeated trials reveal response timing, reliability and condition-dependent structure.

  • Configure or document align by trigger, create rasters, averages and psths and record trigger times, spike times, trial labels.

  • Measure, calculate or compare align by trigger, create rasters, averages and psths using an explicit operational rule.

  • Interpret the result and state why trigger alignment and sample timing must be validated; psth bin width changes interpretation

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Align by trigger, create rasters, averages and PSTHs.

4. Acquire or inspect trigger times, spike times, trial labels.

5. Produce raster/psth figure and trial-exclusion log.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Trigger alignment and sample timing must be validated; PSTH bin width changes interpretation.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate align by trigger, create rasters, averages and psths. The reusable output is raster/psth figure and trial-exclusion log.

Sources / connections / priority. Sources: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]. Natural follow-ons: SEN-05, STA-03. Development priority: High-priority extension. Boundary: Trigger alignment and sample timing must be validated; PSTH bin width changes interpretation.

DAT-07 — Correlation, cross-correlation, phase and synchrony

Concept / theme Relationships between signals require lag-aware measures and controls for common drive.
Audience / context Master’s. Formal practical, workshop or modular course. Prior modules: DAT-04, NET-06, EPH-04.
Logistics 60 min; 1–2 per computer; boards: 0–1; software: Jupyter; GUI for acquisition/QC; prepared dataset: Yes: generated earlier or prepared. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Jupyter analysis. Stages: 6–12.
Spikeling relationship 5 — Recorded-dataset analysis; 3 — Hybrid board + Jupyter; 4 — Multi-board implementation; 8 — External computational comparison

Learning outcomes.

  • Identify and predict the principal behaviour described in relationships between signals require lag-aware measures and controls for common drive.

  • Configure or document analyse paired/multi-board traces with correlation, cross-correlation or phase metrics and record two or more time series/spike trains, lags, phase.

  • Measure, calculate or compare analyse paired/multi-board traces with correlation, cross-correlation or phase metrics using an explicit operational rule.

  • Interpret the result and state why correlation and synchrony do not establish direct connectivity or causation

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Analyse paired/multi-board traces with correlation, cross-correlation or phase metrics.

4. Acquire or inspect two or more time series/spike trains, lags, phase.

5. Produce relationship plot with surrogate/control comparison.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Correlation and synchrony do not establish direct connectivity or causation.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate analyse paired/multi-board traces with correlation, cross-correlation or phase metrics. The reusable output is relationship plot with surrogate/control comparison.

Sources / connections / priority. Sources: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]. Natural follow-ons: STA-06, EPI-03. Development priority: Advanced specialised module. Boundary: Correlation and synchrony do not establish direct connectivity or causation.

DAT-08 — Encoding, decoding and reproducible analysis pipeline

Concept / theme A complete notebook can predict stimulus condition while documenting preprocessing and validation.
Audience / context Master’s. Formal practical, workshop or modular course. Prior modules: DAT-05 or DAT-06, STA-02.
Logistics 60 min; 1–2 per computer; boards: 0–1; software: Jupyter; GUI for acquisition/QC; prepared dataset: Yes: generated earlier or prepared. Equipment: Standard USB cable; worksheet/protocol card.
Mode / stages Jupyter analysis. Stages: 6–12.
Spikeling relationship 5 — Recorded-dataset analysis; 8 — External computational comparison

Learning outcomes.

  • Identify and predict the principal behaviour described in a complete notebook can predict stimulus condition while documenting preprocessing and validation.

  • Configure or document build a simple classifier/regression model with train/test separation and record features, labels, predictions, cross-validation metrics.

  • Measure, calculate or compare build a simple classifier/regression model with train/test separation using an explicit operational rule.

  • Interpret the result and state why performance can reflect confounds, leakage or repeated samples; biological coding claims require restraint

Roadmap.

1. Initial prediction or classification.

2. Configure the board, GUI, simulation or dataset and record metadata.

3. Build a simple classifier/regression model with train/test separation.

4. Acquire or inspect features, labels, predictions, cross-validation metrics.

5. Produce re-runnable notebook and model card.

6. Compare conditions or models and justify the chosen measurement.

7. Answer a limitation question: Performance can reflect confounds, leakage or repeated samples; biological coding claims require restraint.

Inputs → outputs. Students receive a configuration/protocol prompt, variable definitions and any required starter data. They manipulate build a simple classifier/regression model with train/test separation. The reusable output is re-runnable notebook and model card.

Sources / connections / priority. Sources: [T5 Chs.1–4; T6 Chs.7,10–11; O1; O5; G4, G9]. Natural follow-ons: MET-06, EPI-03. Development priority: Advanced specialised module. Boundary: Performance can reflect confounds, leakage or repeated samples; biological coding claims require restraint.