# 20.2.5 oa_neuron_information

This is the table hosting the content of a Self-Organized Map calculated from a dataset composed by outliers by the Apsis module OA. Each entry corresponds to parameters estimated for one particular neuron of the map. The prototype BP/RP spectrum for a particular neuron is available in another table: oa_neuron_xp_spectra. See Section 11.3.12 for further details.

Columns description:

All Gaia data processed by the Data Processing and Analysis Consortium comes tagged with a solution identifier. This is a numeric field attached to each table row that can be used to unequivocally identify the version of all the subsystems that were used in the generation of the data as well as the input data used. It is mainly for internal DPAC use but is included in the published data releases to enable end users to examine the provenance of processed data products. To decode a given solution ID visit https://gaia.esac.esa.int/decoder/solnDecoder.jsp

A unique single numerical identifier of the Self-Organized Map.

A unique single numerical identifier of the neuron.

A row index which determines the position of the neuron within the Self-Organized Map lattice, together with neuron_column_index.

A column index which determines the position of the neuron within the Self-Organized Map lattice, together with neuron_row_index.

Number of sources that have been assigned to the neuron according to the squared Euclidean distance between the sources XP spectra and the neuron XP prototype (neuron_oa_dist in table astrophysical_parameters), see Section 11.3.12 for further details.

Astronomical class estimated for the neuron by means of a template matching procedure among a set of reference templates. This field can be null if no template is assigned. See Section 11.3.12 for further details.

centroid_id : Identifier of the Gaia source that minimizes the classification distance to the neuron (long)

Identifier of the Gaia source (gaia_source.source_id) that minimizes the squared Euclidean distance between the source XP spectrum and the neuron XP prototype. See centroid_distance.

centroid_distance : Squared Euclidean distance between the centroid XP spectrum and the neuron XP prototype (float)

Squared Euclidean distance between the XP spectrum of the centroid source and the neuron XP prototype (xp_spectrum_prototype_flux in table oa_neuron_xp_spectra).

template_distance : Squared Euclidean distance between the reference XP template and the neuron XP prototype (float)

Squared Euclidean distance between the XP spectrum of the reference template (xp_spectrum_template_flux in table oa_neuron_xp_spectra) and the neuron XP prototype (xp_spectrum_prototype_flux in table oa_neuron_xp_spectra). Thie field can be null if no template is assigned. See Section 11.3.12 for further details.

Mean $G$ magnitude value for those sources that belong to the neuron. The mean $G$ magnitude (phot_g_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

g_mag_std_dev : Standard deviation of $G$ values for the sources that belong to the neuron (float, Magnitude[mag])

Standard deviation of $G$ magnitude values for those sources that belong to the neuron. The mean $G$ magnitude (phot_g_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

Minimum $G$ value for those sources that belong to the neuron. The mean $G$ magnitude (phot_g_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

Maximum $G$ value for those sources that belong to the neuron. The mean $G$ magnitude (phot_g_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

bp_mag_mean : Mean ${G}_{\mathrm{BP}}$ value for the sources that belong to the neuron (float, Magnitude[mag])

Mean ${G}_{\mathrm{BP}}$ value for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{BP}}$ magnitude (phot_bp_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

bp_mag_std_dev : Standard deviation of ${G}_{\mathrm{BP}}$ values for the sources that belong to the neuron (float, Magnitude[mag])

Standard deviation of ${G}_{\mathrm{BP}}$ values for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{BP}}$ magnitude (phot_bp_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

bp_mag_min : Minimum value of ${G}_{\mathrm{BP}}$ for the sources that belong to the neuron (float, Magnitude[mag])

Minimum value of ${G}_{\mathrm{BP}}$ for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{BP}}$ magnitude (phot_bp_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

bp_mag_max : Maximum value of ${G}_{\mathrm{BP}}$ for the sources that belong to the neuron (float, Magnitude[mag])

Maximum value of ${G}_{\mathrm{BP}}$ for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{BP}}$ magnitude (phot_bp_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

rp_mag_mean : Mean ${G}_{\mathrm{RP}}$ value for the sources that belong to the neuron neuron (float, Magnitude[mag])

Mean ${G}_{\mathrm{RP}}$ value for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{RP}}$ magnitude (phot_rp_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

rp_mag_std_dev : Standard deviation of ${G}_{\mathrm{RP}}$ values for the sources that belong to the neuron (float, Magnitude[mag])

Standard deviation of ${G}_{\mathrm{RP}}$ values for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{RP}}$ magnitude (phot_rp_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

rp_mag_min : Minimum value of ${G}_{\mathrm{RP}}$ for the sources that belong to the neuron (float, Magnitude[mag])

Minimum value of ${G}_{\mathrm{RP}}$ for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{RP}}$ magnitude (phot_rp_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

rp_mag_max : Maximum value of ${G}_{\mathrm{RP}}$ for the sources that belong to the neuron (float, Magnitude[mag])

Maximum value of ${G}_{\mathrm{RP}}$ for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{RP}}$ magnitude (phot_rp_mean_mag in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

pm_ra_mean : Mean value of the proper motion in right ascension for the sources that belong to the neuron (float, Angular Velocity[mas yr${}^{-1}$])

Mean value of the proper motion in right ascension direction for those sources that belong to the neuron. The proper motion in right ascension (gaia_source.pmra) for each source is used, and if such a value is not finite, then it is not taken into account.

pm_ra_std_dev : Standard deviation of the proper motion in right ascension for the sources that belong to the neuron (float, Angular Velocity[mas yr${}^{-1}$])

Standard deviation of the proper motion in right ascension direction for those sources that belong to the neuron. The proper motion in right ascension (gaia_source.pmra) for each source is used, and if such a value is not finite, then it is not taken into account.

pm_ra_min : Minimum value of the proper motion in right ascension for the sources that belong to the neuron (float, Angular Velocity[mas yr${}^{-1}$])

Minimum value of the proper motion in right ascension direction for those sources that belong to the neuron. The proper motion in right ascension (gaia_source.pmra) for each source is used, and if such a value is not finite, then it is not taken into account.

pm_ra_max : Maximum value of the proper motion in right ascension for the sources that belong to the neuron (float, Angular Velocity[mas yr${}^{-1}$])

Maximum value of the proper motion in right ascension direction for those sources that belong to the neuron. The proper motion in right ascension (gaia_source.pmra) for each source is used, and if such a value is not finite, then it is not taken into account.

pm_dec_mean : Mean value of the proper motion in declination for the sources that belong to the neuron (float, Angular Velocity[mas yr${}^{-1}$])

Mean value of the proper motion in declination direction for those sources that belong to the neuron. The proper motion in declination (gaia_source.pmdec) for each source is used, and if such a value is not finite, then it is not taken into account.

pm_dec_std_dev : Standard deviation of the proper motion in declination for the sources that belong to the neuron (float, Angular Velocity[mas yr${}^{-1}$])

Standard deviation of the proper motion in declination direction for those sources that belong to the neuron. The proper motion in declination (gaia_source.pmdec) for each source is used, and if such a value is not finite, then it is not taken into account.

pm_dec_min : Minimum value of the proper motion in declination for the sources that belong to the neuron (float, Angular Velocity[mas yr${}^{-1}$])

Minimum value of the proper motion in declination direction for those sources that belong to the neuron. The proper motion in declination (gaia_source.pmdec) for each source is used, and if such a value is not finite, then it is not taken into account.

pm_dec_max : Maximum value of the proper motion in declination for the sources that belong to the neuron (float, Angular Velocity[mas yr${}^{-1}$])

Maximum value of the proper motion in declination direction for those sources that belong to the neuron. The proper motion in declination (gaia_source.pmdec) for each source is used, and if such a value is not finite, then it is not taken into account.

Mean parallax value for those sources that belong to the neuron. The parallax (gaia_source.parallax) for each source is used, and if such a value is not finite, then it is not taken into account.

parallax_std_dev : Standard deviation of the parallax values for the sources that belong to the neuron (float, Angle[mas])

Standard deviation of the parallax values for those sources that belong to the neuron. The parallax (gaia_source.parallax) for each source is used, and if such a value is not finite, then it is not taken into account.

Minimum parallax value for those sources that belong to the neuron. The parallax (gaia_source.parallax) for each source is used, and if such a value is not finite, then it is not taken into account.

Maximum parallax value for those sources that belong to the neuron. The parallax (gaia_source.parallax) for each source is used, and if such a value is not finite, then it is not taken into account.

gal_latitude_mean : Mean galactic latitude for the sources that belong to the neuron (float, Angle[deg])

Mean value of the galactic latitude for those sources that belong to the neuron. The galactic latitude (gaia_source.b) for each source is used, and if such a value is not finite, then it is not taken into account.

gal_latitude_std_dev : Standard deviation of the galactic latitude values for the sources that belong to the neuron (float, Angle[deg])

Standard deviation of the galactic latitude values for those sources that belong to the neuron. The galactic latitude (gaia_source.b) for each source is used, and if such a value is not finite, then it is not taken into account.

gal_latitude_min : Minimum galactic latitude for the sources that belong to the neuron (float, Angle[deg])

Minimum value of the galactic latitude for those sources that belong to the neuron. The galactic latitude (gaia_source.b) for each source is used, and if such a value is not finite, then it is not taken into account.

gal_latitude_max : Maximum galactic latitude for the sources that belong to the neuron (float, Angle[deg])

Maximum value of the galactic latitude for those sources that belong to the neuron. The galactic latitude (gaia_source.b) for each source is used, and if such a value is not finite, then it is not taken into account.

intra_neuron_distance_mean : Mean value of the squared Euclidean distance between each of the XP sources in the neuron and the neuron prototype (float)

Mean value of the squared Euclidean distance between each of the XP spectra of the sources that belong to the neuron and the XP prototype of the neuron (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

intra_neuron_distance_std_dev : Standard deviation of the squared Euclidean distance between each of the XP sources in the neuron and the neuron prototype (float)

Standard deviation of the squared Euclidean distance between each of the XP spectra of the sources that belong to the neuron and the XP prototype of the neuron (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

intra_neuron_distance_min : Minimum squared Euclidean distance between each of the XP sources in the neuron and the neuron prototype (float)

Minimum value of the Squared Euclidean distance between each of the XP spectra of the sources that belong to the neuron and the XP prototype of the neuron (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

intra_neuron_distance_max : Maximum squared Euclidean distance between each of the XP sources in the neuron and the neuron prototype (float)

Maximum value of the Squared Euclidean distance between each of the XP spectra of the sources that belong to the neuron and the XP prototype of the neuron (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

inter_neuron_distance_mean : Mean value of the squared Euclidean distance between the neuron XP prototype and the XP prototypes of its immediate neighbours (float)

Mean value of the squared Euclidean distance between the neuron XP prototype (xp_spectrum_prototype_flux in table oa_neuron_xp_spectra) and the XP prototypes of its immediate neighbours. See Section 11.3.12 for further details.

inter_neuron_distance_std_dev : Standard deviation of the squared Euclidean distance between the neuron XP prototype and the XP prototypes of its immediate neighbours (float)

Standard deviation of the squared Euclidean distance between the neuron XP prototype (xp_spectrum_prototype_flux in table oa_neuron_xp_spectra) and the XP prototypes of its immediate neighbours. See Section 11.3.12 for further details.

inter_neuron_distance_min : Minimum value of the squared Euclidean distance between the neuron XP prototype and the XP prototypes of its immediate neighbours (float)

Minimum value of the squared Euclidean distance between the neuron XP prototype (xp_spectrum_prototype_flux in table oa_neuron_xp_spectra) and the XP prototypes of its immediate neighbours. See Section 11.3.12 for further details.

inter_neuron_distance_max : Maximum value of the squared Euclidean distance between the neuron XP prototype and the XP prototypes of its immediate neighbours (float)

Maximum value of the squared Euclidean distance between the neuron XP prototype (xp_spectrum_prototype_flux in table oa_neuron_xp_spectra) and the XP prototypes of its immediate neighbours. See Section 11.3.12 for further details.

Name of the template that has been assigned to the neuron. Templates were built from observed XP spectra of representative astronomical classes and allow to assign an astronomical type or class label to the objects populating a neuron. This field can be null if no template is assigned. See Section 11.3.12 for further details.

25${}^{th}$ percentile value for the intra-neuron distance distribution, which includes the 25% of the worst classified objects. That is, the 25% of objects with the highest squared Euclidean distances to their neuron XP prototype (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

50${}^{th}$ percentile value for the intra-neuron distance distribution, which includes the 50% of the worst classified objects. That is, the 50% of objects with the highest squared Euclidean distances to their neuron XP prototype (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

68${}^{th}$ percentile value for the intra-neuron distance distribution, which includes the 68% of the worst classified objects. That is, the 68% of objects with the highest squared Euclidean distances to their neuron XP prototype (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

75${}^{th}$ percentile value for the intra-neuron distance distribution, which includes the 75% of the worst classified objects. That is, the 75% of objects with the highest squared Euclidean distances to their neuron XP prototype (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

95${}^{th}$ percentile value for the intra-neuron distance distribution, which includes the 95% of the worst classified objects. That is, the 95% of objects with the highest squared Euclidean distances to their neuron XP prototype (oa_neuron_xp_spectra.xp_spectrum_prototype_flux). See Section 11.3.12 for further details.

FWHM is given by the distance between two points on a neuron squared Euclidean distance distribution at which the function reaches half its maximum value. Those neurons showing low values for this quantity contain objects displaying similar spectra, and can be considered as neurons with a good clustering quality. See Section 11.3.12 for further details.

Skewness is the third standardized moment of a distribution. It measures the distortion or asymmetry on a neuron squared Euclidean distance distribution. If the skewness value is positive, then the function is shifted to the left, and if its value is negative to the right, in both cases the function is said to be skewed. See Section 11.3.12 for further details.

Kurtosis is the ratio of the fourth moment ${\mu}^{4}$ to the square of the variance. It is a measure of the heaviness of the tails of the neuron squared Euclidean distance distribution. Those better quality neurons will take a value clearly greater than zero and, therefore, the distribution of distances will tend to be a focused distribution. See Section 11.3.12 for further details.

Difference between the 75${}^{th}$ (distance_percentile75) and 25${}^{th}$ (distance_percentile25) percentile values for the intra-neuron distance distribution. Low values indicate a focused distance distribution. See Section 11.3.12 for further details.

FWHM (distance_fwhm) is given by the distance between two points on a neuron squared Euclidean distance distribution at which the function reaches half its maximum value. We calculate it as a normalized value, by dividing the FWHM mean value obtained for the 10% of neurons with the lowest values in the SOM map by the FWHM in the neuron (distance_fwhm). Those neurons showing values near or above one for this quantity contain objects displaying similar spectra, and can be considered as neurons with a good clustering quality. See Section 11.3.12 for further details.

quality_category : Quality category assigned to the neuron, where 0 corresponds to the most homogeneous neurons and 6 to the most heterogeneous ones (short)

In order to provide a unified interpretation of the goodness of the clustering in a neuron, a general index has been calculated that takes into account the results obtained for the three intra-neuron distance distribution measurements: FWHM (distance_fwhm), skewness (distance_skew) and kurtosis (distance_kurtosis). To do so, the following percentile values of the three indices have been set: ${10}^{th}$, ${32}^{th}$, ${50}^{th}$, ${75}^{th}$, ${90}^{th}$ and ${95}^{th}$. For each of the indices it is possible to obtain seven quality categories depending on the interquartile area where the measures are grouped, and then assign the neuron a global index named quality category (QC). Quality category can take any of the following values: 0 in case all three indices are in percentile level ${95}^{th}$, value 1 if they are in level ${90}^{th}$, and so on to finish with level 6 when all quality indices are outside percentile ${10}^{th}$. Quality category 0 corresponds to the most homogeneous neurons and 6 to the most heterogeneous ones. If no category could be assigned then it takes value -1. See Section 11.3.12 for further details.

bp_transits_mean : Mean value of the number of BP transits for the sources that belong to the neuron (float)

Mean value of the number of BP transits for those sources that belong to the neuron. The number of BP transits (gaia_source.phot_bp_n_obs) for each source is used, and if such a value is not finite, then it is not taken into account.

bp_transits_std_dev : Standard deviation of the number of BP transits for the sources that belong to the neuron (float)

Standard deviation of the number of BP transits for those sources that belong to the neuron. The number of BP transits (gaia_source.phot_bp_n_obs) for each source is used, and if such a value is not finite, then it is not taken into account.

bp_transits_min : Minimum value of the number of BP transits for the sources that belong to the neuron (float)

Minimum value of the number of BP transits for those sources that belong to the neuron. The number of BP transits (gaia_source.phot_bp_n_obs) for each source is used, and if such a value is not finite, then it is not taken into account.

bp_transits_max : Maximum value of the number of BP transits for the sources that belong to the neuron (float)

Maximum value of the number of BP transits for those sources that belong to the neuron. The number of BP transits (gaia_source.phot_bp_n_obs) for each source is used, and if such a value is not finite, then it is not taken into account.

rp_transits_mean : Mean value of the number of RP transits for the sources that belong to the neuron (float)

Mean value of the number of RP transits for those sources that belong to the neuron. The number of RP transits (gaia_source.phot_rp_n_obs) for each source is used, and if such a value is not finite, then it is not taken into account.

rp_transits_std_dev : Standard deviation of the number of RP transits for the sources that belong to the neuron (float)

Standard deviation of the number of RP transits for those sources that belong to the neuron. The number of RP transits (gaia_source.phot_rp_n_obs) for each source is used, and if such a value is not finite, then it is not taken into account.

rp_transits_min : Minimum value of the number of RP transits for the sources that belong to the neuron (float)

Minimum value of the number of RP transits for those sources that belong to the neuron. The number of RP transits (gaia_source.phot_rp_n_obs) for each source is used, and if such a value is not finite, then it is not taken into account.

rp_transits_max : Maximum value of the number of RP transits for the sources that belong to the neuron (float)

Maximum value of the number of RP transits for those sources that belong to the neuron. The number of RP transits (gaia_source.phot_rp_n_obs) for each source is used, and if such a value is not finite, then it is not taken into account.

ruwe_mean : Mean value of the renormalised unit weight error for the sources that belong to the neuron (float)

Mean value of the renormalised unit weight error for those sources that belong to the neuron. The renormalised unit weight error (gaia_source.ruwe) for each source is used, and if such a value is not finite, then it is not taken into account.

ruwe_std_dev : Standard deviation of the renormalised unit weight error for the sources that belong to the neuron (float)

Standard deviation of the renormalised unit weight error for those sources that belong to the neuron. The renormalised unit weight error (gaia_source.ruwe) for each source is used, and if such a value is not finite, then it is not taken into account.

ruwe_min : Minimum value of the renormalised unit weight error for the sources that belong to the neuron (float)

Minimum value of the renormalised unit weight error for those sources that belong to the neuron. The renormalised unit weight error (gaia_source.ruwe) for each source is used, and if such a value is not finite, then it is not taken into account.

ruwe_max : Maximum value of the renormalised unit weight error for the sources that belong to the neuron (float)

Maximum value of the renormalised unit weight error for those sources that belong to the neuron. The renormalised unit weight error (gaia_source.ruwe) for each source is used, and if such a value is not finite, then it is not taken into account.

bprp_mean_flux_excess_mean : Mean value of the BP/RP flux excess for the sources that belong to the neuron (float)

Mean value of the flux excess for those sources that belong to the neuron. The BP/RP excess factor (phot_bp_rp_excess_factor in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

bprp_mean_flux_excess_std_dev : Standard deviation of the BP/RP flux excess for the sources that belong to the neuron (float)

Standard deviation of the flux excess values for those sources that belong to the neuron. The BP/RP excess factor (phot_bp_rp_excess_factor in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

bprp_mean_flux_excess_min : Minimum value of the BP/RP flux excess for the sources that belong to the neuron (float)

Minimum value of the flux excess for those sources that belong to the neuron. The BP/RP excess factor (phot_bp_rp_excess_factor in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

bprp_mean_flux_excess_max : Maximum value of the BP/RP flux excess for the sources that belong to the neuron (float)

Maximum value of the flux excess for those sources that belong to the neuron. The BP/RP excess factor (phot_bp_rp_excess_factor in table gaia_source) for each source is used, and if such a value is not finite, then it is not taken into account.

bprp_colour_mean : Mean value of the ${G}_{\mathrm{BP}}-{G}_{\mathrm{RP}}$ colour for the sources that belong to the neuron (float, Magnitude[mag])

Mean value of the ${G}_{\mathrm{BP}}-{G}_{\mathrm{RP}}$ colour for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{BP}}$ magnitude (gaia_source.phot_bp_mean_mag) and mean ${G}_{\mathrm{RP}}$ magnitude (gaia_source.phot_rp_mean_mag) for each source are used, and if such values are not finite, then it is not taken into account.

bprp_colour_std_dev : Standard deviation of the ${G}_{\mathrm{BP}}-{G}_{\mathrm{RP}}$ colour for the sources that belong to the neuron (float, Magnitude[mag])

Standard deviation of the ${G}_{\mathrm{BP}}-{G}_{\mathrm{RP}}$ colour for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{BP}}$ magnitude (gaia_source.phot_bp_mean_mag) and mean ${G}_{\mathrm{RP}}$ magnitude (gaia_source.phot_rp_mean_mag) for each source are used, and if such values are not finite, then it is not taken into account.

bprp_colour_min : Minimum value of the ${G}_{\mathrm{BP}}-{G}_{\mathrm{RP}}$ colour for the sources that belong to the neuron (float, Magnitude[mag])

Minimum value of the ${G}_{\mathrm{BP}}-{G}_{\mathrm{RP}}$ colour for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{BP}}$ magnitude (gaia_source.phot_bp_mean_mag) and mean ${G}_{\mathrm{RP}}$ magnitude (gaia_source.phot_rp_mean_mag) for each source are used, and if such values are not finite, then it is not taken into account.

bprp_colour_max : Maximum value of the ${G}_{\mathrm{BP}}-{G}_{\mathrm{RP}}$ colour for the sources that belong to the neuron (float, Magnitude[mag])

Maximum value of the ${G}_{\mathrm{BP}}-{G}_{\mathrm{RP}}$ colour for those sources that belong to the neuron. The integrated mean ${G}_{\mathrm{BP}}$ magnitude (gaia_source.phot_bp_mean_mag) and mean ${G}_{\mathrm{RP}}$ magnitude (gaia_source.phot_rp_mean_mag) for each source are used, and if such values are not finite, then it is not taken into account.