Article Figures & Data
Figures
- Figure 1.
Schematic illustration of prototypes. A, Electronic components, where the three selected sensors are shown in the INPUTS box: LDR, photograph emitter/photosensor (photoelectric), and the capacitive sensor (with a 65-mm stainless steel nozzle). An Arduino MEGA was used as an analog-digital signal converter. The shields are shown on the right: an SD card module to record the collected data, a real-time clock module (DS3231) and an LCD 16 × 2 to show, licks counting in real time. B–D, Prototypes’ schematic showing the lick cabin for each sensor, the mouse positioning, the liquid bottle and the sensors. B, LDR prototype sensor was positioned at one side of the lick cabin. C, The photoelectric prototype sensor had the photo emitter aligned to the photoreceiver on the lick cabin. D, The capacitive prototype sensor was created by the connection between the stainless-steel nozzle to an analog input of the Arduino (black wire), and the platform was connected to the ground (Gnd, green wire). Created with BioRender. For electronic schematic diagram, please see Extended Data Figure 1-1.
- Figure 2.
A, Designed acrylic cage dimensions for the lickometer. B, Removable wall with the cabins where drinking sippers are positioned. C, Lateral view. D, Front view.
- Figure 3.
A, Acrylic box designed to store the electronic circuit. B, Top view. C, Top view of the open circuit box. D, Front view shows the cables that connect the sensors with the electronic circuit.
- Figure 4.
Experimental design of the in silico validation protocol with emulated signals. Digital pulses emulating licks were emitted through the digital ports of the microcontroller ESP8266 at frequencies of 5 and 10 Hz. In total, we have built eight cages, each cage has two bottles with sensors that we referred to as A and B. The validation consisted of three phases, in which the number of licks was simulated by alternating each bottle. Phase 1: the pulses initially were sent only from one output A (blue square) to simulate licks from the lickometer A (blue) of one house cage, then two A outputs until simultaneously eight outputs were sending pulses from lickometers A to simulate simultaneous licks from the eight house cages on the lickometers A of each cage. Phase 2: the signals were sent only from the sensor connected to the lickometer B (orange), repeating Phase 1 for this output. Phase 3: pulses were sent from both lickometers A and B, alternately. Thus, signals were sent from one output A (blue) followed by one output B (orange) until simultaneously eight house cages send A and B signals. Created with BioRender.
- Figure 5.
Illustration of one set of lickometer house cages with a detailed view of its parts. A, Schematic of the central system that transmits data from the sensors to the computer wirelessly. The central system consists of an ESP8266, a real-time clock (RTC) and a multiplexer. One central is connected to 16 lickometers from eight boxes. B, One actual set of eight house cages. C, Draft of the final house cage highlighting its parts. Both stainless steel nozzles are connected to an input of the central system (A: lickometer with blue wire and B: lickometer with orange wire), and the platform is connected to the ground (green wire). Created with BioRender.
- Figure 6.
Results of recall and precision values for two prototypes. A, Diagram of the selected prototypes for validation: capacitive and photoelectric sensors connected to the Analog/Digital (A/D) converter. B, Descriptive results of the precision of the photoelectric and capacitive prototypes validation for female and male mice. Each graphical bar is the mean ± standard deviation of the video evaluated by different researchers. C, Descriptive results of the recall of the photoelectric and capacitive prototypes validation for female and male mice. Each graphical bar is the mean ± standard deviation of the video evaluated by different researchers. D, Graphical visualization of results of recall and precision for capacitive and photoelectric prototypes (median ± 95% confidence interval). An independent t test detected a significant difference in precision between prototypes (*p < 0.05), being higher with slower dispersion for the capacitive sensor. Recall remained similar for both prototypes, without significant differences.
- Figure 7.
Results of reading and transmission errors obtained with the in silico validation of 5 and 10 Hz emulated frequencies for increasing number of cages (channels) receiving the signal (1–8). Box plots represent the mean and the whiskers 9th–95th percentile. Output A is represented in blue and output B is represented in orange. We describe the general range of the errors in this legend using mean ± standard deviation. A, Reading error in Phase 1 (output A) at 5 Hz ranged around 0.0422 ± 0.035%, and at 10 Hz, around 0.0559 ± 0.0424%. B, Reading error in Phase 2 (output B) at 5 Hz ranged around 0.0173 ± 0.0817%, and at 10 Hz, around 0.056 ± 0.0415%. C, Reading error in Phase 3. Considering the 5-Hz frequency, output A shows a reading error ranging around 0.0466 ± 0.0521%, and output B, around 0.0375 ± 0.059%. At 10 Hz, output A shows a reading error ranging around 0.0594 ± 0.054%, and output B, around 0.0679 ± 0.0495%. D, Transmission error in Phase 1 (output A) at 5 Hz ranged around −0.422 ± 0.494%, and at 10 Hz, around −0.484 ± 1.073%. E, Transmission error in Phase 2 (output B) at 5 Hz ranged around −0.661 ± 0.61%, and at 10 Hz, around −1.187 ± 1.618%. F, Transmission error in Phase 3. Considering the 5-Hz frequency, output A shows a transmission error ranging around −0.898 ± 1.312%, and output B, around −1.013 ± 1.405%. At 10 Hz, output A shows a transmission error ranging around −0.441 ± 0.479%, and output B, around −0.537 ± 0.589%. Horizontal lines represent statistically significant differences of each error and each frequency in relation to the amount of channels, calculated by the Bonferroni post hoc analysis of an independent-samples Kruskal–Wallis test, p < 0.05.
- Figure 8.
Results of the in vivo validation of the final lickometer system. A, Schematic representation of the one final cage containing two lickometers coupled to the central unit. B, Dispersion graph of the consumed liquid (difference of the bottle weight at the end of the experiment compared with the beginning of the experiment in grams - g) and total number of licks for each mouse in each session in the lickometer system. Spearman correlation revealed a strong correlation between the variables (p < 0.05). C, Dispersion graph of the consumed liquid (g) and total number of licks for each mouse in each session in the lickometer system separated by different types of liquids: water and 10% ethanol. Correlation analysis revealed strong correlation independently of the liquid presented to the animals (p < 0.05). D, Example trace of the raw acquired data demonstrating the detailed microstructure and timestamp licks that can be collected by our device. High peaks indicate licks events.
Tables
- Table 1
List of materials used to build the full system of eight house cages and the components of the electronic central unit
Materials Quantity Unit cost Total cost Estimated total cost (US$) ESP8266 1 R$ 30.00 R$ 30.00 US$ 6.00 Module RTC 1 R$ 25.00 R$ 25.00 US$ 5.00 Module Mux 74 hc4067 1 R$ 15.00 R$ 15.00 US$ 3.00 Pin bar 1 × 40 6 R$ 1,50 R$ 9.00 US$ 1.80 Switched mode power supply 5V 1A 1 R$ 12.00 R$ 12.00 US$ 2.40 PowerSupply connector/jack P4 female 5.5 x 2.1 mm 1 R$ 0.60 R$ 0.60 US$ 0.12 Female BNC connectors 32 R$ 2.20 R$ 70.40 US$ 14.08 Male BNC connector RG59 75R 32 R$ 4.80 R$ 153.60 US$ 30.72 Coax cable RG59 750HM 96% 16 R$ 3.00 R$ 48.00 US$ 9.60 Banana jack 4 mm 8 R$ 1.30 R$ 10.40 US$ 2.08 Cable 0.25 mm 8 R$ 1.00 R$ 8.00 US$ 1.60 Female pin B67 8 R$ 7.00 R$ 56.00 US$ 11.20 Male pin P22 8 R$ 0.16 R$ 1.28 US$ 0.26 FBU 323101 terminal 16 R$ 0.30 R$ 4.80 US$ 0.96 Resistor 100K 2 R$ 0.05 R$ 0,10 US$ 0.02 General expenses (glue, resin…) 1 R$ 100.00 R$ 100.00 US$ 20.00 Printed circuit board for grounding 8 R$ 30.00 R$ 240.00 US$ 48.00 Printed circuit board for wireless central 1 R$ 48.68 R$ 48.68 US$ 9.74 Acrylic box for circuit 1 R$ 200.00 R$ 200.00 US$ 40.00 Acrylic cage 8 R$ 220.00 R$ 1760.00 US$ 352.00 Welding the nozzle to the connector 16 R$ 20.00 R$ 36.00 US$ 7.20 Total cost R$ 2828.86 US$ 565.78 The values were considered with the Brazilian currency (real, R$) in February 2021 and converted to the United States dollar. The dollar exchange rate was considered on February 23, 2022.
Extended DATAFigure 1-1
Extended figure shows the electronic schematic diagram of the three built prototypes: (A) photoelectric, (B) LDR, and (C) capacitive sensor. Download Figure 1-1, TIF file.
In this issue
