Visiomode: An open-source platform for building rodent touchscreen-based behavioral assays

Authors: Constantinos Eleftheriou, Thomas Clarke, V. Poon, Marie Zechner, Ian Duguid (University of Edinburgh: Simons Initiative for the Developing Brain; Centre for Discovery Brain Sciences; Roslin Institute). Journal: Journal of Neuroscience Methods 386:109779 (2023 Feb 15). DOI 10.1016/j.jneumeth.2022.109779. PMCID PMC10375831 (open access, CC BY 4.0).

Abstract

Background

Touchscreen-based behavioral assays provide a robust method for assessing cognitive behavior in rodents, offering great flexibility and translational potential. The development of touchscreen assays presents a significant programming and mechanical engineering challenge, where commercial solutions can be prohibitively expensive and open-source solutions are underdeveloped, with limited adaptability.

New method

Here, we present Visiomode (www.visiomode.org), an open-source platform for building rodent touchscreen-based behavioral tasks. Visiomode leverages the inherent flexibility of touchscreens to offer a simple yet adaptable software and hardware platform. The platform is built on the Raspberry Pi computer combining a web-based interface and powerful plug-in system with an operant chamber that can be adapted to generate a wide range of behavioral tasks.

Results

As a proof of concept, we use Visiomode to build both simple stimulus-response and more complex visual discrimination tasks, showing that mice display rapid sensorimotor learning including switching between different motor responses (i.e., nose poke versus reaching).

Comparison with existing methods

Commercial solutions are the ‘go to’ for rodent touchscreen behaviors, but the associated costs can be prohibitive, limiting their uptake by the wider neuroscience community. While several open-source solutions have been developed, efforts so far have focused on reducing the cost, rather than promoting ease of use and adaptability. Visiomode addresses these unmet needs providing a low-cost, extensible platform for creating touchscreen tasks.

Conclusions

Developing an open-source, rapidly scalable and low-cost platform for building touchscreen-based behavioral assays should increase uptake across the science community and accelerate the investigation of cognition, decision-making and sensorimotor behaviors both in health and disease.

Highlights

• • Open-source platform for building rodent touchscreen behavioral tasks.

• • Combines a web-based interface and powerful USB plug-in system.

• • Plug-and-play design enables rapid adaptability and scaling of touchscreen tasks.

• • Novel touchscreen-based 2-AFC forelimb reaching task for rodents.

1. Introduction

Since their introduction to biomedical research, touchscreens have become an increasingly popular tool for assessing cognitive function in rodents ( Bussey et al., 1997, Dumont et al., 2021, Markham et al., 1996). Their appeal lies with their remarkable flexibility, supporting a vast array of visual stimuli coupled with quantifiable motor responses ( Seitz et al., 2021), both of which are necessary for designing tasks to investigate complex cognitive processes such as category learning ( Broschard et al., 2021, Kim et al., 2018), spatial attention ( Haddad et al., 2021), cognitive flexibility ( Groman et al., 2012), and visual perception ( Markham et al., 1996). Their use has transformed studies of neurological disorders by providing a sensitive assay of sensorimotor behaviors ( Arulsamy et al., 2019, Copping et al., 2017, Leach and Crawley, 2018, Leach et al., 2016, Morton et al., 2006, Norris et al., 2019, Yang et al., 2015), revealing subtle phenotypes that were not detected by more conventional assays ( Van den Broeck et al., 2019, Zeleznikow-Johnston et al., 2018). This sensitive readout of changes in behavior holds great translational promise ( Talpos and Steckler, 2013), where tasks designed for animals can be directly translated to human subjects ( Chow et al., 2020, Hvoslef-Eide et al., 2015, Nithianantharajah et al., 2015). Despite their increasing popularity, the use of touchscreen-based behaviors is somewhat limited in rodent research. Uptake has been hampered either by the prohibitive up-front costs of commercial systems or the considerable ‘in-house’ development required to create bespoke touchscreen-based behaviors ( Dumont et al., 2021).

Commercially available touchscreen behavioral arenas provide researchers with a simple turnkey solution requiring minimal setup time. These systems have dominated the touchscreen landscape in biomedicine over the past two decades ( Arulsamy et al., 2019, Brasted et al., 2002, Brigman et al., 2010, Bussey et al., 1998, Bussey et al., 2008, Delotterie et al., 2014, Glover et al., 2020, Haddad et al., 2021, Heath et al., 2019, Odland et al., 2021, Piantadosi et al., 2019, Stirman et al., 2016, Talpos et al., 2008), and have played an important role in popularizing their use in rodent research ( Dumont et al., 2021). However, the prohibitive costs associated with commercial systems (i.e., > 10,000 USD) provides a rate limiting step for their widespread adoption ( Dumont et al., 2021). In contrast, developing touchscreen tasks ’in-house’ is a particularly challenging programming and engineering problem. While most traditional open-field ( Hall and Ballachey, 1932) or operant chamber ( Skinner, 1938) tasks can be implemented with a simple microcontroller device ( Akam et al., 2022), the introduction of a touchscreen interface requires complex hardware and software integration to control the generation and display of graphics, as well as registering behavioral interactions with the screen. Utilizing graphics libraries available on most Operating Systems (e.g., Microsoft Windows, Linux and MacOS) requires extensive programming knowledge ( Kessenich et al., 2016), and while open-source initiatives such as PsychoPy greatly simplify the task of generating visual stimuli ( Peirce, 2007), they still require significant development to be adapted for touchscreen tasks ( Seitz et al., 2021). This is further complicated by the choice of touchscreen hardware, where heterogeneity in compact touchscreen systems results in variable touch sensitivities, requiring the developer to test and validate a range of screens before final implementation ( Dumont et al., 2021).

In our view, an open-source, community-driven touchscreen solution would solve both problems by distributing the development effort across multiple research groups, while also reducing overall costs ( Fortunato and Galassi, 2021, Freeman, 2015). Open-science initiatives have continued to grow in the past few years, with projects like MouseBytes ( Beraldo et al., 2019) and the advice sharing platform touchscreencognition.org ( Dumont et al., 2021). To date, no open-source community-based solution exists. While open-source touchscreen-based operant chambers have been developed ( Gurley, 2019, O’Leary et al., 2018, Pineno, 2014), this has not led to increased uptake due a lack of code availability and the primary focus being on reducing costs, rather than enhancing the user experience, scalability, adaptability, and ease of use.

To address these unmet needs, we have develo

Key points for the wiki

• Raspberry Pi-based platform: web interface + plugin system + adaptable operant chamber; generates a wide range of touchscreen tasks.
• Demonstrated tasks: simple stimulus-response and 2-AFC visual discrimination (nose poke and reaching variants); mice show rapid sensorimotor learning and can switch motor responses.
• Low-cost, extensible, ease-of-use focus (vs commercial chambers and earlier cost-focused open-source efforts).
• Data export supports JSON, HDF5 and NWB formats; real-time analysis.