Baijiu classification method based on brain magnetic resonance imaging

Abstract

The invention relates to the field of baijiu classification, in particular to a baijiu classification method based on brain magnetic resonance imaging, which comprises the following steps of: S101, sniffing a baijiu sample by a subject for 5-10 minutes; s102, obtaining brain functional magnetic resonance imaging data; s103, obtaining corresponding resting state brain network data; s104, extracting the sub-network data therein; s105, obtaining a relative intensity value of activation or suppression of the sub-network; and S106, based on the obtained relative intensity value, classifying the aroma types of the white spirit sample. According to the baijiu classification method provided by the invention, the aroma types of the baijiu samples can be accurately classified without being interfered by other factors, and an objective basis is provided for the aspects of baijiu aroma type classification, quality evaluation and the like.

Description

Technical Field

[0001] The invention relates to the field of liquor classification, and in particular to a liquor classification method based on brain magnetic resonance imaging. Background Art

[0002] Chinese liquor is made from starch-rich grains as the main raw material. It is an alcoholic beverage made by using Daqu, Xiaoqu, Fuqu and yeast as saccharification and fermentation agents, using solid state (some types of wine are semi-solid or liquid) fermentation, distillation, storage and blending. The brewing process and ingredients of liquor are complicated. Even with the same raw materials, fermentation agents and fermentation environment, the flavor, taste and quality of liquor may vary.

[0003] At present, the quality evaluation of most liquors in my country still relies on traditional sensory evaluation and has not been replaced by analytical instruments (such as chromatographs and mass spectrometers), which is significantly different from other food identification evaluations. The main reason for still using sensory evaluation is that the liquor itself has its own unique aroma (such as strong aroma, light aroma, and sauce aroma), and even the flavors of different liquor production batches are different, and these flavors are important factors that affect consumers' choice of liquor. Common analytical instruments such as chromatographs and mass spectrometers cannot analyze the flavor of liquor and can only simply give quantitative data. Therefore, the current classification method of liquor flavor still relies on sensory evaluation. However, this sensory evaluation is based on manual subjective evaluation, and the evaluation of different people is often subjective. Moreover, its evaluation is easily affected by uncontrollable factors such as limited accuracy of sensory discrimination and inconsistent evaluation indicators, making it difficult for different people to reach a unified conclusion on the classification of liquor flavors.

[0004] Regarding the use of olfactory stimulation to identify aroma samples, Chinese patent application CN102680644A discloses a method for identifying aroma samples that trigger rewards through dopaminergic pathways, which uses task-state functional magnetic resonance imaging, that is, making the subject smell during the magnetic resonance scanning process to provide feedback on the activation of specific brain regions. This method actually reflects a one-to-one stimulus-brain region mapping relationship in a biological sense, that is, it can only reflect whether a certain aroma sample can activate the dopaminergic pathway, and cannot achieve the classification of aroma samples. Summary of the invention

[0005] The present invention provides a liquor classification method based on brain magnetic resonance imaging, characterized in that the method classifies the flavor type of liquor samples, comprising the following steps:

[0006] S101 allows the subject to smell the liquor sample for 5-10 minutes;

[0007] S102, after the subject finishes sniffing, performing a brain functional magnetic resonance imaging scan on the subject to obtain brain functional magnetic resonance imaging data;

[0008] S103 preprocesses the brain functional magnetic resonance imaging data and obtains corresponding resting-state brain network data based on a standard brain template;

[0009] S104 extracts sub-network data based on the resting-state brain network data. The subnetworks include the dorsal attention network, the anterior default mode network, the posterior default mode network, and the basal ganglia network;

[0010] S105 uses the resting state brain network in the standard brain template as a mask image, Extracting the BOLD signal intensity corresponding to the sub-network data, and then obtaining the relative intensity value of activation or inhibition of the sub-network;

[0011] S106 classifies the flavor type of the liquor sample based on the relative strength value obtained, wherein when the relative strength value of the basal ganglia network is greater than a set threshold, the liquor sample is classified as phoenix-flavor type; when the relative strength value of the anterior default network is greater than a set threshold, the liquor sample is classified as light-flavor type; when the relative strength value of the posterior default network is greater than a set threshold, the liquor sample is classified as sauce-flavor type; when the relative strength value of the dorsal attention network is greater than a set threshold, the liquor sample is classified as strong-flavor type.

[0012] In some embodiments, the number of subjects is at least 5.

[0013] In some embodiments, the relative intensity value is the average of the relative intensity values ​​of all subjects.

[0014] In some embodiments, the preprocessing methods include: time layer correction, realignment, redundancy regression, normalization and smoothing.

[0015] In some embodiments, the threshold is set to 8-10.

[0016] In some embodiments, the time interval between step S101 and step S102 comprises 10-15 minutes.

[0017] In some embodiments, the step S105 specifically includes:

[0018] S201 combines the BOLD signal in the sub-network data and its corresponding time series into first image data;

[0019] S202 regresses the time series matrix information and the 3D space series matrix information in the first image data respectively;

[0020] S203 extracts the first image data after the time series matrix information is regressed, Obtaining second image data; and using the resting-state brain network in the standard brain template as a mask image, extracting the BOLD signal value in the mask image in the second image data, that is, the BOLD signal intensity corresponding to the sub-network data;

[0021] S204 calculates the average value of the BOLD signal value in the mask image. The relative strength value of the activation or inhibition of the sub-network is obtained.

[0022] In some embodiments, the threshold is set to 10.

[0023] In some embodiments, the number of subjects is 10-20.

[0024] Compared with the prior art, the beneficial effects of the present invention include at least the following aspects:

[0025] At present, the main method for classifying the flavor of liquor is sensory evaluation, which requires well-trained and experienced sensory evaluators, and the training process is time-consuming and less cost-effective. It may take several months to half a year to train a sensory evaluator. In addition, even for well-trained sensory evaluators, the results are easily interfered by various factors, such as exposure duration, fatigue, and even containers. Common analytical instruments such as chromatographs and mass spectrometers cannot analyze the flavor of liquor and can only provide quantitative data.

[0026] The prior art has not reported the use of activation or inhibition of resting state networks to achieve objective classification of liquor flavors. The brain functional magnetic resonance imaging (fMRI) used in the present invention is a neuroimaging method for non-invasive analysis of brain activity. It uses the physical basis that deoxyhemoglobin is more paramagnetic than oxyhemoglobin, and completes imaging by detecting the contrast between deoxyhemoglobin and oxyhemoglobin (i.e., BOLD signal). And the present invention uses resting state fMRI (rs-fMRI), that is, the subject performs magnetic resonance scanning without a specific task (without systematic thinking or trying not to think about the problem) to reflect the spontaneous neural activity of the human brain to different flavor liquors. That is, the present invention uses the different stimuli caused by different flavor liquors to change the cerebral blood flow in the brain area to affect the activation or inhibition of the brain network, which can be captured by the functional magnetic resonance scanning used by the present invention. When a certain brain area is active, the blood flow in the area will increase, resulting in an increase in the oxygen content in the surrounding blood. Compared with the unused brain area, this oxygenated blood has a different effect on the magnetic field, thereby producing signal differences in magnetic resonance imaging. Furthermore, the present invention utilizes the signal differences generated by each subnetwork (including the basal ganglia network, the anterior default network, the posterior default network, and the dorsal attention network), that is, the present invention utilizes the specific activation results of specific brain networks exhibited by the stimulation of the brain by the subjects smelling different flavors of liquor, to achieve objective classification of liquor flavors. Compared with the prior art, the liquor classification method provided by the present invention can accurately classify the flavor of liquor samples, has high repeatability, and is not easily interfered by other factors, and provides an objective basis for the classification of liquor flavors (such as the identification of the flavor of newly developed liquors), quality assessment (such as adulteration, imitation), etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, The following is a brief introduction to the drawings required for use in the embodiments or prior art descriptions. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale. Obviously, the drawings described below are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without paying creative labor.

[0028] Figure 1 This is a diagram showing the activation of the basal ganglia network induced by olfactory stimulation in response to olfactory stimulation, which is different from other types of liquor.

[0029] Figure 2 This is a diagram showing the activation of the default network before olfactory stimulation induced by the light-flavor liquor compared to other flavor liquors;

[0030] Figure 3This is a diagram showing the activation of the default mode network after olfactory stimulation, which distinguishes the sauce-flavor liquor from other flavor liquors;

[0031] Figure 4 This is a diagram showing the activation of the dorsal attention network induced by olfactory stimulation in response to the difference between Luzhou-flavor liquor and other types of liquor;

[0032] Figure 5 Schematic diagram of a brain network in a low-activation / no-activation state;

[0033] Figure 6 The present invention provides a flow chart of the liquor classification method. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0035] Herein "and / or" includes any and all combinations of one or more of the associated listed items.

[0036] Herein, "plurality" means two or more than two, ie, it includes two, three, four, five, etc.

[0037] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the sentence “Comprising a…” defined element does not exclude the existence of other identical elements in the process, method, article or device comprising the element.

[0038] As used in this specification, the term "about" typically means + / - 5% of the stated value, more typically + / - 4% of the stated value, more typically + / - 3% of the stated value, more typically + / - 2% of the stated value, even more typically + / - 1% of the stated value, and even more typically + / - 0.5% of the stated value.

[0039] In this specification, some embodiments may be disclosed in a format of being in a certain range. It should be understood that such descriptions of "being in a certain range" are merely for convenience and brevity, and should not be interpreted as a rigid limitation on the disclosed range. Therefore, the description of a range should be considered to have specifically disclosed all possible sub-ranges and independent numerical values ​​within this range. For example, the description of a range of 1-6 should be considered to have specifically disclosed sub-ranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6, etc., as well as individual numbers within this range, such as 1, 2, 3, 4, 5 and 6. The above rules apply regardless of the breadth of the range.

[0040] Embodiment 1

[0041] Subjects: 20-100 healthy subjects were recruited. Inclusion and exclusion criteria included:

[0042] Inclusion criteria: (1) aged 18-35 years; (2) no obvious abnormalities in head MRI scan; (3) good physical condition and no history of major diseases; (4) voluntary participation and signing of informed consent.

[0043] Exclusion criteria: (1) patients with systemic diseases or mental illnesses; (2) patients with contraindications to MRI scanning such as claustrophobia or metal implants; (3) patients who are unable to cooperate due to cognitive dysfunction or other reasons; (4) pregnant or lactating women.

[0044] Sniffing and fMRI scanning:

[0045] 10 minutes before functional magnetic resonance imaging (Siemens 3.0T prisma) scanning, Healthy subjects were asked to smell a certain type of liquor sample for 5 minutes, avoiding interference from other irritating odors. Functional magnetic resonance imaging (about 30 minutes) was then performed. After the scan, the subjects were asked to rest for 7 days, and then the above steps were repeated with liquor samples of different flavors until all liquor samples were smelled and scanned with functional magnetic resonance imaging.

[0046] Scan sequence parameters:

[0047] i) T1w-sMRI (T1 weighted image - structural magnetic resonance imaging): voxel size: 0.8×0.8×0.8mm; Repetition time (TR): 2400ms; Echo time (TE): 2.11ms; Inversion time (TI): 1000ms; FoV read: 256mm; slice thickness: 0.8mm; basic resolution: 320; slabs: 1; dist.factor: 50%; position: isocenter; Orientation: sagittal plane; phase encoding direction (phase enc.dir): A>>P;

[0048] ii) SpinEchoFieldMap_RL-fMRI (functional magnetic resonance imaging): voxel size: 2.1×2.1×2.1mm; repetition time: 8000ms; Echo time: 66ms;

[0049] iii)SpinEchoFieldMap_LR-fMRI: Voxel size: 2.1×2.1×2.1mm; repetition time: 8000ms; echo time: 66ms;

[0050] iv) rfMRI (resting state functional magnetic resonance imaging)_REST_RL: Voxel size: 2.1×2.1×2.1mm; repetition time: 700ms; echo time: 37.80ms; Layer: 64; Phase encoding direction: R>>L; FoV reading: 210mm; FoV phase (FoVphase): 84$; Layer thickness: 2.1mm;

[0051] v) rfMRI_REST_LR: voxel size: 2.1×2.1×2.1 mm; repetition time: 700 ms; echo time: 37.80 ms; slices: 64; phase encoding direction: L>>R; FoV reading: 210mm; FoV phase: 84$; Slice thickness: 2.1mm.

[0052] Post-processing:

[0053] The data obtained from the above operations were post-processed to explore the differences in brain function of subjects after olfactory stimulation of liquor samples with different flavors, so as to conduct liquor classification based on functional magnetic resonance imaging.

[0054] FSL was used for data preprocessing based on the HCP pipeline, including slicetiming, realignment, nuisance regression, normalization, and smoothing. Principal component analysis (ICA) was performed on the preprocessed data using fsl-melodic to explore 20-120 component segmentations and determine brain networks based on standard templates to determine the resting-state brain networks in the templates. Independent component maps related to white matter, cerebrospinal fluid, or physiological noise were classified and excluded from further study. The remaining brain networks were determined as resting-state brain networks. The anatomical locations of specific brain regions in the above brain networks were determined based on the Harvard-Oxford cortical and subcortical structural atlas.

[0055] Extract the relative strength value of brain network activation or inhibition. The specific method is as follows:

[0056] A. Integrate the subject's BOLD signal and its time series and merge them into a 4D nii file;

[0057] B. Regress the time series matrix information and 3D space series matrix information of the above 4D files respectively;

[0058] C. Extract the nii file after the regression of the above time series matrix information, use the resting state brain network in the generated template as the mask file, and extract the BOLD signal value within the mask;

[0059] D. Calculate the average value of the BOLD signal value in the above mask using a high-order algorithm. Obtain the relative strength value of brain network activation or inhibition (continuous variable).

[0060] Use network activation values ​​to determine the type of liquor:

[0061] The network activation intensity values ​​in the above templates were extracted, and the differences in brain network activation intensity values ​​of different types of liquor were compared. The random permutation test in FSL was used to test the inter-group differences, and 5000 permutations and threshold-free clustering enhancement (TFCE) were performed for multiple comparison correction. When the p value was less than 0.05, the effect was considered statistically significant. Through statistical analysis and comparison, the specific network activation degree (average activation intensity ≥ 10) and inhibition degree (-1 ≥mean activation intensity ≥ -5).

[0062] Embodiment 2

[0063] This example uses the method of Example 1 to analyze the degree of brain network activation or inhibition of 20 subjects in response to four common flavor liquors (Luzhou-flavor liquor, Maotai-flavor liquor, Qing-flavor liquor, and Feng-flavor liquor). The results are as follows: Figure 1-Figure 4 shown.

[0064] Among them, the specific liquors tested are Wuliangye (strong aroma type), Langjiu (sauce aroma type), Red Star Erguotou (light aroma type), and Xifeng Liquor (Phoenix aroma type).

[0065] See also Figure 1 , Figure 1 The left picture shows the sagittal image of the basal ganglia network activation caused by Fengxiang-type liquor, the middle picture shows the coronal image of the basal ganglia network activation caused by Fengxiang-type liquor, and the right picture shows the axial image of the basal ganglia network activation caused by Fengxiang-type liquor. The results showed that in each brain network, what distinguished Fengxiang-type liquor from other flavors in the manual classification was the high intensity activation of the basal ganglia network induced by olfactory stimulation (i.e., the average relative intensity value of the basal ganglia network was 11.7), while other brain networks were in a low activation / no activation state (i.e., the relative intensity values ​​of other brain networks were 1-5). For a schematic diagram of a brain network in a low activation / no activation state, see Figure 5 .

[0066] See also Figure 2 , Figure 2 The left picture shows the sagittal image of the pre-default mode network activation caused by light-fragrance liquor, the middle picture shows the coronal image of the pre-default mode network activation caused by light-fragrance liquor, and the right picture shows the axial image of the pre-default mode network activation caused by light-fragrance liquor. The results show that in each brain network, the light-fragrance liquor in the manual classification is different from other flavors in that the olfactory stimulation induces a high intensity activation of the pre-default network (that is, the average relative intensity value of the pre-default network is 12.3), while other brain networks are in a low activation / no activation state.

[0067] See also Figure 3 , Figure 3 The left picture shows the sagittal image of the activation of the posterior default mode network caused by Maotai-flavor liquor, the middle picture shows the coronal image of the activation of the posterior default mode network caused by Maotai-flavor liquor, and the right picture shows the axial image of the activation of the posterior default mode network caused by Maotai-flavor liquor. The results show that in each brain network, the Maotai-flavor liquor in the manual classification is different from other flavors in that the posterior default mode network induced by olfactory stimulation has a high intensity activation (that is, the average relative intensity value of the posterior default network is 12.1), while other brain networks are in a low activation / no activation state.

[0068] See also Figure 4 , Figure 4The left picture shows the sagittal image of the dorsal attention network activation caused by Luzhou-flavor liquor, the middle picture shows the coronal image of the dorsal attention network activation caused by Luzhou-flavor liquor, and the right picture shows the axial image of the dorsal attention network activation caused by Luzhou-flavor liquor. The results show that in each brain network, what distinguishes Luzhou-flavor liquor from other flavors in the manual classification is the high intensity activation of the dorsal attention network induced by olfactory stimulation (that is, the average relative intensity value of the dorsal attention network is 13.6), while other brain networks are in a low activation / no activation state. Figure 1-Figure 5 Generated by MRIcron software.

[0069] As a verification, this embodiment then analyzes the degree of brain network activation or inhibition of the four types of liquor in another 20 subjects, and the results are similar to the above (P<0.05), that is, in each brain network, the Fengxiang type liquor in the artificial classification is distinguished from other types by the high intensity activation of the basal ganglia network induced by olfactory stimulation (that is, the average relative intensity value of the basal ganglia network is 11.3); the light-flavor liquor in the artificial classification is distinguished from other types by the high intensity activation of the pre-default network induced by olfactory stimulation (that is, the average relative intensity value of the pre-default network is 12.5); the sauce-flavor liquor in the artificial classification is distinguished from other types by the high intensity activation of the post-default mode network induced by olfactory stimulation (that is, the average relative intensity value of the post-default network is 12.4); the strong-flavor liquor in the artificial classification is distinguished from other types by the high intensity activation of the dorsal attention network induced by olfactory stimulation (that is, the average relative intensity value of the dorsal attention network is 13.8).

[0070] The above results show that different types of liquor can specifically and highly activate the specific resting-state brain networks of the subjects. Therefore, the flavor of liquor can be objectively classified through resting state functional magnetic resonance imaging (fMRI) based on the specific brain network activation or inhibition degree caused by the olfactory stimulation of different types of liquor.

[0071] Embodiment 3

[0072] See also Figure 6 In order to more objectively classify the flavor of liquor samples, this embodiment provides a liquor classification method based on brain magnetic resonance imaging, comprising the following steps:

[0073] S101 asked the subjects to smell the liquor sample for 5-10 min;

[0074] In some embodiments, the number of subjects is at least 5.

[0075] In some embodiments, the number of subjects is 10-20.

[0076] S102 performs a functional brain magnetic resonance imaging scan on the subject after the subject finishes sniffing to obtain functional brain magnetic resonance imaging data; functional brain magnetic resonance imaging is a neuroimaging method for non-invasive analysis of brain activity, which uses the physical basis that deoxyhemoglobin is more paramagnetic than oxyhemoglobin to complete imaging by detecting the contrast between deoxyhemoglobin and oxyhemoglobin (i.e., BOLD signal).

[0077] Task state and resting state are two states of the brain. The former mainly refers to the state of the brain when it performs specific tasks such as memory, recognition and movement, while the resting state refers to the state when the brain does not perform specific cognitive tasks, remains quiet, relaxed and awake. The present invention uses resting state fMRI technology to achieve objective classification of liquor flavors by studying the activation intensity of the resting state network of the brain caused by the olfactory stimulation of different types of liquor on the subjects.

[0078] In some embodiments, the time interval between step S101 and step S102 comprises 10-15 minutes.

[0079] S103 preprocesses the brain functional magnetic resonance imaging data and obtains corresponding resting-state brain network data based on a standard brain template;

[0080] In some embodiments, the preprocessing method includes: slice timing correction, realignment, nuisance regression, normalization and smoothing. The above preprocessing helps to eliminate the influence of white matter, cerebrospinal fluid or physiological noise on resting brain network data, thereby obtaining resting brain network data with physiological significance.

[0081] In some embodiments, the standard brain template is a standard meningeal plate based on a region of interest. Specifically, the region of interest can be used to define brain network nodes, and the brain can be divided into several regions of interest according to the standard brain template, for example, the entire brain is divided into several regions of interest based on the Harvard-Oxford cortical and subcortical structural atlas, wherein each region of interest represents a node of the brain network. In some embodiments, the standard brain template is an MNI (Montreal Neuroscience Research Center) template, such as MNI305, Colin27, MNI152, etc. If the brain network nodes are defined using regions of interest, the voxel point set corresponding to each region of interest is determined respectively, and the time series of each voxel point in the preprocessed resting state brain network data is extracted, and finally the average time series of the voxels contained in each region of interest is calculated respectively to obtain the BOLD signal intensity of the corresponding region of interest. In some embodiments, the areas of interest include visual network (VIN), sensorimotor network (SEN), auditory network (AUN), cerebellum network (CEN), dorsal attention network (DAN), ventral attention network (VAN), salience network (SAN), left frontoparietal network (LFPN), right frontoparietal network (RFPN), basal ganglia network, executive control network (ECN) and default mode network (DMN).

[0082] S104 extracts sub-network data based on the resting-state brain network data. The subnetworks include the dorsal attention network, the anterior default mode network, the posterior default mode network and the basal ganglia network. Due to the complex brain structure and different brain region division standards, the processing time and results of the same resting-state brain network data will be affected. Based on the experimental results of the previous embodiments, the present invention focuses on analyzing the data of the above sub-network in the resting-state brain network data, which is highly targeted and reduces the amount of calculation, saving time and labor costs;

[0083] S105 uses the resting state brain network in the standard brain template as a mask image, Extracting the BOLD signal intensity corresponding to the sub-network data, and then obtaining the relative intensity value of activation or inhibition of the sub-network;

[0084] In some embodiments, the relative intensity value is the average of the relative intensity values ​​of all subjects.

[0085] In some embodiments, the step S105 specifically includes:

[0086] S201 combines the BOLD signal in the sub-network data and its corresponding time series into first image data;

[0087] S202 regresses the time series matrix information and the 3D space series matrix information in the first image data respectively;

[0088] S203 extracts the first image data after the time series matrix information is regressed, Obtaining second image data; and using the resting-state brain network in the standard brain template as a mask image, extracting the BOLD signal value in the mask image in the second image data, that is, the BOLD signal intensity corresponding to the sub-network data;

[0089] S204 calculates the average value of the BOLD signal value in the mask image. The relative strength value of the activation or inhibition of the sub-network is obtained.

[0090] S106 classifies the flavor type of the liquor sample based on the relative strength value obtained, wherein when the relative strength value of the basal ganglia network is greater than a set threshold, the liquor sample is classified as phoenix-flavor type; when the relative strength value of the anterior default network is greater than a set threshold, the liquor sample is classified as light-flavor type; when the relative strength value of the posterior default network is greater than a set threshold, the liquor sample is classified as sauce-flavor type; when the relative strength value of the dorsal attention network is greater than a set threshold, the liquor sample is classified as strong-flavor type.

[0091] In some embodiments, the threshold is set to 8-10.

[0092] In some embodiments, the threshold is set to 10.

[0093] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A liquor classification method based on brain magnetic resonance imaging, characterized in that: The method classifies the flavor types of liquor samples, comprising the following steps: S101 allows the subject to smell the liquor sample for 5-10 minutes; S102, after the subject finishes sniffing, performing a brain functional magnetic resonance imaging scan on the subject to obtain brain functional magnetic resonance imaging data; S103 preprocesses the brain functional magnetic resonance imaging data and obtains corresponding resting-state brain network data based on a standard brain template; S104 extracting subnetwork data based on the resting-state brain network data, wherein the subnetworks include a dorsal attention network, an anterior default mode network, a posterior default mode network, and a basal ganglia network; S105 uses the resting state brain network in the standard brain template as a mask image, extracts the BOLD signal intensity corresponding to the sub-network data, and then obtains the relative intensity value of activation or inhibition of the sub-network; S106 classifies the flavor type of the liquor sample based on the relative strength value obtained, wherein when the relative strength value of the basal ganglia network is greater than a set threshold, the liquor sample is classified as phoenix-flavor type; when the relative strength value of the anterior default network is greater than a set threshold, the liquor sample is classified as light-flavor type; when the relative strength value of the posterior default network is greater than a set threshold, the liquor sample is classified as sauce-flavor type; when the relative strength value of the dorsal attention network is greater than a set threshold, the liquor sample is classified as strong-flavor type.

2. The method according to claim 1, characterized in that The number of the subjects is at least 5.

3. The method according to claim 1, characterized in that The relative intensity value is the average of the relative intensity values ​​of all subjects.

4. The method according to claim 1, characterized in that The preprocessing method includes: time layer correction, realignment, redundancy regression, normalization and smoothing.

5. The method according to claim 1, characterized in that The threshold is set to 8-10.

6. The method according to claim 1, characterized in that The time interval between step S101 and step S102 is 10-15 minutes.

7. The method according to claim 1, characterized in that The step S105 specifically includes: S201 combines the BOLD signal in the sub-network data and its corresponding time series into first image data; S202 regresses the time series matrix information and the 3D space series matrix information in the first image data respectively; S203 extracts the first image data after the time series matrix information is regressed to obtain the second image data; and uses the resting state brain network in the standard brain template as a mask image to extract the BOLD signal value in the mask image in the second image data, that is, the BOLD signal intensity corresponding to the sub-network data; S204 calculates the average value of the BOLD signal value in the mask image to obtain a relative intensity value of activation or inhibition of the sub-network.

8. The method according to claim 1, characterized in that The standard brain template is the MNI template.

9. The method according to claim 1, characterized in that The threshold is set to 10.

10. The method according to claim 1, characterized in that The number of the subjects is 10-20.

Images