Optogenetic Experiments

[pauladkisson]

Notes for optogenetic stimulation in optoda_raw_data/closed_loop_behavior.parquet.

[pauladkisson]

Overall this file seems to relate to the experiments in which optogenetic stimulation was delivered during targeted syllables in a closed loop fashion see Figure 3

[pauladkisson]

151 Columns split into semantic groups

[pauladkisson]

Syllable Related Columns

• Same as in #1 EXCEPT
• no labels field AND
• a new syllable_group_legacy field --> probably some old version of the hierarchical syllable clustering
• In order to match feedback_status to target_syllable, one needs to use predicted_syllable (online) rather than predicted_syllable (offline) --> need to include both types of syllable information in the syllable event series

Conclusion: Need to update syllable representation to include both online and offline syllables.

[pauladkisson]

PCs and Likes

• PCs are same as #1 BUT
• Contains 100 "likes" fields (likes00, likes01, ..., likes99)
  • All likes are strictly negative --> log-likelihoods of each syllable?

Conclusion: log-likelihoods as predicted by some model are also too derived to include in the nwb dataset.

[pauladkisson]

Session-related Columns

Same as #1 EXCEPT

• Missing session_total_number, stim_session, and unique_idx
• Has extra experiment_type field = categorical variable with fields:
  • 'excitation'
  • has mouse_ids = {'vta-nacc-ctrl-6', '3442', '3155', '3440', 'snc-dls-ctrl-6', '3441'}
  • unclear what (if any) experiments this refers to
  • 'excitation_photometry'
  • has mouse_ids = {'dlight-chrimson-1', ..., 'dlight-chrimson-6'}
  • some 6-mouse subset of the 8-mouse opto-da + FP experiments
  • 'excitation_pulsed'
  • refers to the set of experiments using 3s-pulsed stim BUT with no FP
  • 'excitation_pulsed_photometry'
  • refers to the set of experiments using 3s-pulsed stim WITH concurrent FP (Extended Data Figure 10l)
  • 'reinforcement'
  • refers to the set of experiments using 20 opto-DA mice + 12 controls WITHOUT FP (Extended Data Figure 9)
  • 'reinforcement_photometry'
  • mouse_ids = {'dlight-chrimson-1', ..., 'dlight-chrimson-6', 'dlight-chrimson-8', 'dlight-chrimson-9'}
  • refers to the 8 mice that received opto-DA stim with concurrent FP (Methods: Stereotaxic surgery for simultaneous photometric recordings and optogenetic stimulation)

Conclusion: For now, focus on experiment_types that are well-understood with example figures to benchmark against ('excitation_pulsed_photometry', 'reinforcement', 'reinforcement photometry'). When we get a chance, we need to ask the Datta Lab people to explain the other experiment_types.

[pauladkisson]

Session Duplication

• The reinforcement_photometry sessions from optoda_raw_data/closed_loop_behavior.parquet are duplicated in dlight_raw_data/dlight_photometry_processed_full.parquet
• The excitation_pulsed_photometry sessions from optoda_raw_data/closed_loop_behavior.parquet are duplicated in dlight_raw_data/3s-pulsed-stim-dataframe.parquet

BUT

• only the dlight_raw_data/* dataframes have fiber photometry information

Conclusion: May need to load both dataframes to get all the relevant information for overlapping sessions.

[pauladkisson]

Subject-related Columns

Same as #1 EXCEPT

• 'subject_name' and 'SubjectName' are missing
• 'area' = brain region of opto-stim = {'snc (axon)', 'ctrl'}
• 'area (pooled)' = NULL --> useless
• 'opsin' = opsin type = {'ctrl', 'chr2', 'chrimson'}
• 'genotype' = area + opsin information BUT specifies different types of controls
  • experiment_type = 'excitation':
  • area = 'ctrl' --> genotype = {'snc-eyfp', 'snc-dls-ctrl', 'vta-nacc-ctrl'}
  • area = 'snc (axon)' --> genotype = {'snc-dls-ai32'}
  • experiment_type = 'excitation_photometry':
  • area = 'snc (axon) --> genotype = 'dls-chrimson-dlight'
  • experiment_type = 'excitation_pulsed':
  • area = 'ctrl' --> genotype = {'snc-dls-ctrl', 'vta-nacc-ctrl'}
  • area = 'snc (axon)' --> genotype = 'snc-dls-ai32'
  • experiment_type = 'excitation_pulsed_photometry':
  • area = 'snc (axon)' --> genotype = 'dls-chrimson-dlight'
  • experiment_type = 'reinforcement'
  • area = 'ctrl' --> genotype = {'snc-dls-ctrl', 'vta-nacc-ctrl', 'ctrl'}
  • area = 'snc (axon)' --> genotype = {'snc-dls-ai32', 'snc-dls-ai32-jrcamp1a'}
  • experiment_type = 'reinforcement_photometry'
  • area = 'snc (axon)' --> genotype = 'dls-chrimson-dlight'
• sex = mouse sex (missing from dlight_raw_data) = {'male', 'female', None}
• exclude = {True, False} = whether or not a baseline session was excluded from dlight analysis (see Stereotaxic surgery for simultaneous photometric recordings and optogenetic stimulation)
• exclude_reason = None --> useless
• cohort = integer = 0-15 --> some kind of cohort that each mouse was in
• mouse_id = regular and organized for experiment_type == 'reinforcement_photometry' | 'excitation_photometry' and , but all over the place with number-only IDs and irregular ids for other experiment_types (See Session-related Columns)

Conclusion: Add relevant metadata fields (experiment_type, area, sex, opsin, exclude, cohort) to session_metadata.yaml AND I need to see if I can track down some of this metadata (ex. sex) for the dlight-only animals.

[pauladkisson]

Trigger-related Columns

• No trigger-related columns in this dataframe --> Maybe it's related to Closed-loop velocity modulation experiments

[pauladkisson]

Kinematic Data Columns

• Identical to #1

[pauladkisson]

OptoDA Columns

• stim_duration = duration of optogenetic stimulation (seconds) = {0.25, 2, 3} (See Methods: Stereotaxic surgery for simultaneous photometric recordings and optogenetic stimulation)
• stim_frequency = frequency of optogenetic stimulation (Hz) = {nan for CW stim; 25 for 3s-stim} (See Methods: Stereotaxic surgery for simultaneous photometric recordings and optogenetic stimulation)
• pulse_width = pulse width of optogenetic stimulation (seconds) = {nan for CW stim; 0.005 for 3s-stim} (See Methods: Stereotaxic surgery for simultaneous photometric recordings and optogenetic stimulation)
• power = power of optogenetic stimulation (mW) = {10} (See Methods: Mice used to assess the influence of endogenous dopamine fluctuations on optogenetic reinforcement)
• indicator = None or 'dlight' depending on if the animal was used in photometry experiments or not --> redundant info
• feedback_status =
  • 1 if feedback optogenetic stim is applied
  • 0 if no optogenetic stim is applied ('catch trials' =~ 25% of target syllable detections)
  • -1 if predicted_syllable != target_syllable OR before/after o-stim
  • feedback_status only changes for 1 point / syllable --> onset signal?
  • See fig_3l, 3m-optoda entropy and scalar effects.ipynb