'RAVE' repository: with entire recording blocks

Loads recording blocks - continuous recording chunks, typically a run of minutes.

Usage

prepare_subject_with_blocks(
  subject,
  electrodes = NULL,
  blocks = NULL,
  reference_name = NULL,
  ...,
  quiet = FALSE,
  repository_id = NULL,
  strict = TRUE
)

prepare_subject_raw_voltage_with_blocks(
  subject,
  electrodes = NULL,
  blocks = NULL,
  reference_name = "noref",
  downsample = NA,
  ...,
  quiet = FALSE,
  repository_id = NULL,
  strict = TRUE
)

prepare_subject_voltage_with_blocks(
  subject,
  electrodes = NULL,
  blocks = NULL,
  reference_name = NULL,
  downsample = NA,
  ...,
  quiet = FALSE,
  repository_id = NULL,
  strict = TRUE
)

prepare_subject_time_frequency_coefficients_with_blocks(
  subject,
  electrodes = NULL,
  blocks = NULL,
  reference_name = NULL,
  ...,
  quiet = FALSE,
  repository_id = NULL,
  strict = TRUE
)

prepare_subject_phase_with_blocks(
  subject,
  electrodes = NULL,
  blocks = NULL,
  reference_name = NULL,
  ...,
  quiet = FALSE,
  repository_id = NULL,
  strict = TRUE
)

prepare_subject_power_with_blocks(
  subject,
  electrodes = NULL,
  blocks = NULL,
  reference_name = NULL,
  ...,
  quiet = FALSE,
  repository_id = NULL,
  strict = TRUE
)

Arguments

subject

'RAVE' subject

electrodes

string or integers indicating electrodes to load

blocks

names of the recording blocks to load, can be queried via subject$blocks

reference_name

name of the reference table

...

passed to RAVESubjectBaseRepository constructor

quiet

see field quiet

repository_id

see field repository_id

strict

whether to check existence of subject before loading data; default is true

downsample

positive integer or NA, indicating whether the signals should be down-sampled during loading, for voltage traces only; default is NA, meaning no down-sampling

Value

A RAVESubjectRecordingBlockRepository instance

Details

prepare_subject_with_blocks does not actually load any signal data. Its existence is simply for backward compatibility. It instantiates a super-class of the rest of methods. Therefore, please refer to the rest of the methods for loading specific data types.

If you do not need to analyze super high-frequency signals, it is recommended to set a proper downsample value to down-sample the signals while loading voltage traces. This helps optimizing the data storage and speed up computation (significantly). For example, suppose you have 200 channels sampled at 30,000 Hz, a 30-minute recording will cost around 80+ gigabyte memory only to store, let along the storage needed to compute analyses and time needed to perform those analyses. Down-sampling the channels helps a lot. If you are mostly interested in signals below 100 Hz, then down-sampling voltage traces to 400 Hz will preserve the frequency components needed, and it takes 1.2 gigabytes to hold the same recording in memory.

Due to the large-data nature of blocks of signals, the repository will prepare cache files for all the channels, allowing users to load the cached data later without needing to reload

Examples

if( has_rave_subject("demo/DemoSubject") ) {


# ---- An use-case example ------------------------------------------------
# Install subject via install_subject("DemoSubject")
subject <- as_rave_subject("demo/DemoSubject")

# list all blocks
subject$blocks

repository <- prepare_subject_voltage_with_blocks(
  subject,
  electrodes = 13:16,
  blocks = "008",
  reference = "default"
)

print(repository)

repository$blocks

# get data
container <- repository$get_container()

# block data
container$`008`
lfp_list <- container$`008`$LFP
channel_sample_rate <- lfp_list$sample_rate

# Even we only load channels 14-16, all the channels are here for
# in case we want to use the cache for future purposes
lfp_list$dimnames$Electrode

# Plot all loaded channels
channel_sel <- lfp_list$dimnames$Electrode %in% c(14, 15, 16)
channel_signals <- lfp_list$data[, channel_sel,
                                 drop = FALSE,
                                 dimnames = FALSE]

ravetools::plot_signals(t(channel_signals),
                        sample_rate = channel_sample_rate,
                        channel_names = 14:16)

# Load channel 14 and plot pwelch
channel_sel <- lfp_list$dimnames$Electrode == 14

channel_signals <- lfp_list$data[, channel_sel,
                                 drop = TRUE,
                                 dimnames = FALSE]

ravetools::diagnose_channel(channel_signals,
                            srate = channel_sample_rate,
                            name = "Channel 14",
                            nclass = 30)

# ---- Use cache ---------------------------------------------------

subject <- as_rave_subject("demo/DemoSubject")

# Lazy-load block 008
repository <- prepare_subject_voltage_with_blocks(
  subject,
  electrodes = 13:16,
  blocks = "008",
  reference = "default",
  lazy_load = TRUE  # <-- trick
)

# Immediately load data with force=FALSE to use cache if exists
repository$mount_data(force = FALSE)

# ---- More examples ---------------------------------------------


subject <- as_rave_subject("demo/DemoSubject")
repository <- prepare_subject_power_with_blocks(
  subject,
  electrodes = 14,
  blocks = "008",
  reference_name = "default"
)

block_008 <- repository$power$`008`$LFP

channel_sel <- block_008$dimnames$Electrode == 14

# Drop electrode margin
power <- block_008$data[, , channel_sel,
                        drop = TRUE, dimnames = FALSE]

# global baseline
power_baselined_t <- 10 * log10(t(power))
power_baselined_t <- power_baselined_t - rowMeans(power_baselined_t)

ravetools::plot_signals(
  power_baselined_t,
  sample_rate = block_008$sample_rate,
  channel_names = block_008$dimnames$Frequency,
  space = 1,
  start_time = 20,
  duration = 30, ylab = "Frequency",
  main = "Channel 14 - Power with Global Baseline (20-50 sec)"
)


}