Binned, single subject example¶

See mkpy docs for explanations of all the commands and files. Click on the command to find it in the glossary and learn more. The glossary links are incuded for each command below

Conda environment¶

always activate a conda environment before starting. This notebook uses mkconda 0.0.11

Python Libraries¶

from pathlib import Path
from mkpy import mkh5
import pandas as pd

Data Paths¶

Set paths and variable names for files later

MKDIG_DIR = Path("../../mkdig")
MKPY_DIR = Path("../../mkpy")

File names¶

Set file names so they can be easily changed and don’t have to retype every time

# h5 continuous file 
h5_f = str(MKPY_DIR / ("stm02.h5"))

# h5 epoch file 
epochtable = str(MKPY_DIR / ("stm02.epochs.h5"))

# subject ID
sub = 'stm02'

# kutas lab data files (from DIG + yhdr that you created)
eeg = MKDIG_DIR / (sub + ".crw")
log = MKDIG_DIR / (sub +".x.log")
yhdr = MKPY_DIR / (sub + ".yhdr")

Build the continuous h5 file¶

mkh5.mkh5: Import the h5 file into a dataframe and reset it (generally you will want to reset the file to start fresh).

myh5 = mkh5.mkh5(h5_f)
myh5.reset_all()

create_mkdata: load in subject and cals to the h5 data structure

The first argument is the data branch you want to add the data to. This can be subject name, or for more complicated files could be a path.

2a) append_mkdata: if your file is split into two for some reason (i.e., separate cals, DIG crash) this is where you would add this command as well

myh5.create_mkdata(sub, eeg, log, yhdr)

calibrate_mkdata: calibrate data; the first line sets the values similar to normerp

If you don’t know what values to use, you can use plotcals to select them before running the calibrate command.

pts, pulse, lo, hi, ccode = 5, 10, -40, 40, 0
myh5.calibrate_mkdata(sub, # specific data group
                    n_points = pts,   # pts to average
                    cal_size = pulse, # uV
                    lo_cursor = lo,   # lo_cursor ms
                    hi_cursor = hi,   # hi_cursor ms
                    cal_ccode= ccode) # condition code

/home/astoermann/.conda/envs/test_jupyter_books_11/lib/python3.6/site-packages/mkpy/mkh5.py:3706: UserWarning: negative event code(s) found for cal condition code 0 -16384
  warnings.warn(msg)

Found cals in /stm02/dblock_5
Calibrating block /stm02/dblock_0 of 6: (148224,)  

Calibrating block /stm02/dblock_1 of 6: (146944,)

Calibrating block /stm02/dblock_2 of 6: (148224,)

Calibrating block /stm02/dblock_3 of 6: (271104,)

Calibrating block /stm02/dblock_4 of 6: (122624,)

Calibrating block /stm02/dblock_5 of 6: (19200,)

Note

The warning above is telling you there is a negative event code in the cals (the pause) that wasn’t expected. This warning is fine to get. Sometimes pauses happen in the cals i.e, pausing before the cal pulses are completely done.

Find event codes in the data¶

get_event_table: Label the events in the data, but don’t pull out epochs yet. Capture this info in a data frame.

For this simple example we will use the .ytbl format for the code map (see Code Maps ytbl example)

event_table = myh5.get_event_table(MKPY_DIR / ("stmath.ytbl"))
## Look at the event table you just made
event_table.head(312)

searching codes in: stm02/dblock_0
searching codes in: stm02/dblock_1
searching codes in: stm02/dblock_2
searching codes in: stm02/dblock_3

searching codes in: stm02/dblock_4
searching codes in: stm02/dblock_5

	data_group	dblock_path	dblock_tick_idx	dblock_ticks	crw_ticks	raw_evcodes	log_evcodes	log_ccodes	log_flags	epoch_match_tick_delta	...	match_str	anchor_code	match_code	anchor_tick	match_tick	anchor_tick_delta	is_anchor	condition	regexp	STthreat
0	stm02	stm02/dblock_0	66	13565	13565	11025	11025	1	0	0	...	11025	11025	11025	13565	13565	0	True	CorrectNoST	(#11[0-9]{3})	NoST
1	stm02	stm02/dblock_0	108	21958	21958	11003	11003	1	32	0	...	11003	11003	11003	21958	21958	0	True	CorrectNoST	(#11[0-9]{3})	NoST
2	stm02	stm02/dblock_0	192	38634	38634	11078	11078	1	0	0	...	11078	11078	11078	38634	38634	0	True	CorrectNoST	(#11[0-9]{3})	NoST
3	stm02	stm02/dblock_0	234	47106	47106	11088	11088	1	0	0	...	11088	11088	11088	47106	47106	0	True	CorrectNoST	(#11[0-9]{3})	NoST
4	stm02	stm02/dblock_0	248	50191	50191	11039	11039	1	0	0	...	11039	11039	11039	50191	50191	0	True	CorrectNoST	(#11[0-9]{3})	NoST
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
306	stm02	stm02/dblock_4	568	95246	809742	16061	16061	1	0	0	...	16061	16061	16061	95246	95246	0	True	RelatedST	(#16[0-9]{3})	ST
307	stm02	stm02/dblock_4	624	104618	819114	16143	16143	1	0	0	...	16143	16143	16143	104618	104618	0	True	RelatedST	(#16[0-9]{3})	ST
308	stm02	stm02/dblock_4	638	106928	821424	16141	16141	1	0	0	...	16141	16141	16141	106928	106928	0	True	RelatedST	(#16[0-9]{3})	ST
309	stm02	stm02/dblock_4	680	114040	828536	16040	16040	1	0	0	...	16040	16040	16040	114040	114040	0	True	RelatedST	(#16[0-9]{3})	ST
310	stm02	stm02/dblock_4	722	120715	835211	16044	16044	1	0	0	...	16044	16044	16044	120715	120715	0	True	RelatedST	(#16[0-9]{3})	ST

311 rows × 26 columns

Set epochs¶

set_epochs: Pull out epochs in the data. Does not save them yet.

First argument is the name of the epochs table
Second argument is the event table you made (not the code map)
Third and fourth argument is the window you want the epoct to have around 0

myh5.set_epochs('stmath', event_table, -100, 900) 

Sanitizing event table data types for mkh5 epochs table ...

Export epochs¶

export_epochs: save the epoched data to use later. Can be in h5, pdh5 (pandas), or feather format.

First argument is the epochs table
Second argument is the file name to save to (set above in the file names section)
Third argument is what you want to save as

myh5.export_epochs('stmath', epochtable, file_format='h5')

Converting .crw to .h5