Dataset Open Access
Pouzat, Christophe; Mazor, Ofer; Laurent, Gilles
This depository contains raw data from 14 experiments performed on adult locusts. The data are contained in HDF5 files (http://www.hdfgroup.org/HDF5/). They are stored as compressed integers coded on 16 bits as they came out of the A/D card. Recording details can be found in Pouzat, Mazor and Laurent (2002) Using noise signature to optimize spike-sorting and to assess neuronal classification quality. Journal of Neuroscience Methods 122: 43-57 (a pre-print version is available: http://xtof.perso.math.cnrs.fr/pdf/Pouzat+:2002.pdf). Each data file is subdivided in Groups corresponding the type of acquisition performed: one or several epochs of spontaneous activity recording; repetitive stimulation with a given odor. Each group is made of one (if say a single epoch of 60 seconds of spontaneous recording was made) or several (if say 100 stimulation with Citral were made) (sub-)groups containing the data of all the channels that were recorded during that epoch. Each of these sub-groups is made of 4 to 16 data sets: 1 dimensional arrays containing the raw data recorded from one of the 16 channels of our probe (made of 4 tetrodes) during a single acquisition epoch. All channels were sampled at 15 kHz. Each data file has attributes (metadata) README and LabBook. The first, README contains a shortened version of the present text; the second, LabBook contains a transcript of the lab book corresponding to the experiment. Most groups have a log_file_content attribute. This attribute contains a copy a text file that was automatically generated during data acquisition. Some recording details can be found there as well as the precise time and data of each recorded epoch. The data were kept for 14 years on CDs and about a third of the recordings got lost because of CD corruption! What's left still make 15 GBytes of data after compression: a substantial amount. This CD corruption explains why some groups don't have a log_file_content attribute: it was on a corrupted CD.
Of the 14 experiments, 12 contain antennal lobe (the first olfactory relay of insects) recordings, 1 contains antennal lobe and alpha lobe recordings and 1 contains only alpha lobe recordings. Here the alpha lobe location should be taken with a little bit of caution since the latter is not as easy to locate as the antennal lobe in the locust. All data files start with the locust prefix, followed by the experiment year, month and date, like locust20000214.hdf5 an experiment performed on February 14 2000. Due to file size restriction on Zenodo, two experiments are split into two data files: locust20010124b_part1.hdf5 and locust20010124b_part2.hdf5 as well as locust20010214_part1.hdf5 and locust20010214_part2.hdf5. On two dates, two different experiments were performed: locust20010124a.hdf5 and locust20010124b_part1.hdf5 / locust20010124b_part2.hdf5 as well as locust20010208a.hdf5 and locust20010208b.hdf5. When a stimulation was applied, the following code is used: Odor name / Number of stimulation / Inter-stimulation interval / duration before the odor pulse / odor pulse duration / post pulse duration [odor dilution when several dilutions were used]. All times are in seconds.
A very brief description at the group level of the files content follows (see the LabBook attribute of each individual file for details):