A liquid-delivery device that provides precise reward control for neurophysiological and behavioral experiments

https://doi.org/10.1016/j.jneumeth.2005.07.012Get rights and content

Abstract

Behavioral neurophysiology and other kinds of behavioral research often involve the delivery of liquid rewards to experimental subjects performing some kind of operant task. Available systems use gravity or pumps to deliver these fluids, but such methods are poorly suited to moment-to-moment control of the volume, timing, and type of fluid delivered. The design described here overcomes these limitations using an electronic control unit, a pressurized reservoir unit, and an electronically controlled solenoid. The control unit monitors reservoir pressure and provides precisely timed solenoid activation signals. It also stores calibration tables and does on-the-fly interpolation to support computer-controlled delivery calibrated directly in milliliters. The reservoir provides pressurized liquid to a solenoid mounted near the subject. Multiple solenoids, each supplied by a separate reservoir unit and control unit, can be stacked in close proximity to allow instantaneous selection of which liquid reward is delivered. The precision of droplet delivery was verified by weighing discharged droplets on a commercial analytical balance.

Introduction

The value of a primary reinforcer depends upon the quantity of the reinforcer, the relative timing between the reinforcer and the recipient's response, and the recipient's preference for the reinforcer, among other factors. Whereas many behavioral experiments are relatively insensitive to small variations in the features of reward delivery, certain classes of experiments require careful control of all three features just mentioned: quantity, timing, and preference. Described here is an instrument that can provide the needed precision in experiments where liquids are used as a primary reinforcer. Using this device, the amount, timing, and type of liquid presented as primary reinforcement are under direct, moment-to-moment experimental control.

Although suitable for other behavioral applications, the liquid-delivery device was developed for primate behavioral neurophysiological experiments. Neurophysiologists have identified a number of brain areas with neurons sensitive to the timing, amount, or quality of a primary reinforcer. Some neurons, for example, are highly sensitive to the time of reinforcement delivery (Fiorillo et al., 2003), others on its quantity (Cromwell and Schultz, 2003), and yet others on the level of preference for the reward (Tremblay and Schultz, 1999, Watanabe et al., 2002). Further investigation of these findings will benefit from greater control of rewards than traditional liquid-delivery systems provide.

The following design criteria guided the development of a new precision liquid-delivery device:

  • delivery of water and more viscous liquids (e.g., apple sauce)

  • accurate and consistent delivery of volumes as small as 0.05 ml

  • computer-control of delivery volumes

  • delivery initiation within 10 ms of an external event

  • delivery volume independent of suction on the liquid tube

  • capacity for moment-to-moment selection of different liquids

  • rapid, simple readjustment when the liquid type is changed

  • minimum reservoir capacity of 350 ml

  • easy maintenance

Existing liquid-delivery systems do not meet these requirements. Simple gravity-driven liquid dispensers (e.g., Crist Instrument model RLD-E1) do not work well with thicker liquids, are often inconsistent in their delivery volumes, and are sensitive to suction by the animal. When thicker liquids are used, different viscosity liquids tend to be delivered at very different rates. Peristaltic pump systems (e.g., Crist Instrument model 5-RLD-E4) can provide better performance with thicker liquids, but the pump is generally difficult to position near the animal. Motor-operated syringe pumps have the same positioning limitation. In both designs a long liquid-delivery tube is required. Because the tube acts as a reservoir, its length, internal diameter, and distensibility can impact the consistency of volume delivery for viscous solutions or varying loads (e.g., suction by the animal). Stiff, narrow tubing can reduce delivery variations but is more difficult to handle and may clog more easily. Motor-operated syringe pumps and peristaltic pumps may be slow at initiating liquid-delivery, depending upon the implementation.

The delivery device described here circumvents the shortcomings of earlier designs by combining three important design features. First, a pressurized bottle holds the fluid to make its delivery independent of gravity effects, suction or moderate resistance. Second, a small solenoid placed in close proximity to the delivery tube gates the flow. Using a pressurized supply line and placing the solenoid close to the delivery point obviates the need for long delivery tubes and assures the quick release of liquid. Third, a microcontroller controls delivery by calculating delivery parameters on-the-fly based on stored calibration tables. The microcontroller also provides multiple modes of operation, tracks liquid usage, and provides a sophisticated user interface.

Section snippets

General description and design features

The new liquid-delivery device has three major components: an electronic control unit, a reservoir unit, and a solenoid (Fig. 1). The solenoid is an electrically activated valve that is mounted near the animal's mouth. It releases pressurized liquid when activated by electrical pulses. The reservoir unit supports the liquid bottle (reservoir) and regulates air pressure to the bottle. The reservoir unit also provides digitized pressure data to the control unit. The control unit houses a power

Performance

The key performance requirement is precision in the volume of liquid delivered. A number of factors can influence the consistency of liquid-delivery. The control unit must provide accurately-timed pulses to the solenoid. Based on measurements using a storage oscilloscope (Tektronix TDS 210), pulse duration variation (jitter) was either within 1.5% (above 40.0 ms) or less than 200 μs (below 40.0 ms). Variability can also be introduced by the dominance of turbulence at short valve opening times or

Discussion

This report describes a new device for precision delivery of liquid reward for behavioral neurophysiology experiments. Flexible and precise reward delivery enables new opportunities for conducting neurophysiological investigations into reward-related systems in the brain and for other behavioral applications involving precise control over the volume, timing, and type of primary reinforcement. Negative liquid reinforcers could also be delivered with precision. The liquid-delivery system has a

Acknowledgement

This research was supported by the Intramural Research Program of the NIH, NIMH.

References (5)

There are more references available in the full text version of this article.

Cited by (0)

View full text