Vol.:(0123456789)
1 3
Cognitive Therapy and Research
https://doi.org/10.1007/s10608-018-9923-8
ORIGINAL ARTICLE
Personalizing Affective Stimuli Using a Recommender Algorithm:
An Example with Threatening Words for Trauma Exposed Populations
Andrea N. Niles
1,2
· Aoife O’Donovan
1,2
© This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2018
Abstract
Experimental paradigms used in affective and clinical science often use stimuli such as images, scenarios, videos, or words
to elicit emotional responses in study participants. Choosing appropriate stimuli that are highly evocative is essential to the
study of emotional processes in both healthy and clinical populations. Selecting one set of stimuli that will be relevant for
all subjects can be challenging because not every person responds the same way to a given stimulus. Machine learning can
facilitate the personalization of such stimuli. The current study applied a novel statistical approach called a recommender
algorithm to the selection of highly threatening words for a trauma-exposed population (N = 837). Participants rated 513
threatening words, and we trained a user–user collaborative filtering recommender algorithm. The algorithm uses similarities
between individuals to predict ratings for unrated words. We compared threat ratings for algorithm-based word selection to
a random word set, a word set previously used in research, and trauma-specific word sets. Algorithm-selected personalized
words were more threatening compared to non-personalized words with large effects (ds = 2.10–2.92). Recommender algo-
rithms can automate the personalization of stimuli from a large pool of possible stimuli to maximize emotional reactivity in
research paradigms. These methods also hold potential for the personalization of behavioral treatments administered remotely
where a provider is not available to tailor an intervention to the individual. The word personalization algorithm is available
for use online (https ://threa t-word-predi ctor.herok uapp.com/).
Keywords Anxiety · Posttraumatic stress disorder · Emotion · Affect · Recommender algorithm · Trauma · Personalization ·
Personalized medicine
Introduction
Laboratory and computer-based paradigms used in affective
and clinical science frequently employ evocative images,
scenarios, videos, and words to elicit an emotional response
in study participants. For example, the International Affec-
tive Picture System (IAPS; Lang et al. 1999), a picture
set that includes positive, neutral, and negative images, is
widely used in clinical and affective science and has been
cited over 3130 times as of March 7, 2018. Researchers typi-
cally select a set of stimuli that they believe will be relevant
for the phenomenon they are studying (e.g. fear, disgust,
excitement, sadness) and hope that the stimuli do in fact reli-
ability elicit the intended emotion in most or all study par-
ticipants. However, although a stimulus may strongly elicit a
certain emotion on average, the experience of that stimulus
will differ depending on the individual. For example, ratings
of the IAPS images have a mean standard deviation of 1.6
for valence and 2.2 for arousal on a 1–9 scale, meaning that
it is quite common for participants to rate the stimuli 1.5–2
points above or below the mean rating. Thus, methods to
select personalized stimuli for study participants can facili-
tate greater precision when activating emotional responses
in the lab and reduce noise resulting from individual differ-
ences in responses to stimuli. Further, such methods could
be used to personalize the content of behavioral treatments
administered via Internet and mobile applications. In the
current study, we developed and tested an algorithm to select
emotionally evocative stimuli that are tailored to an indi-
vidual participant. In the current demonstration, we select
* Andrea N. Niles
andrea.niles@ucsf.edu
1
Department of Psychiatry and Weill Institute
for Neurosciences, University of California, San Francisco,
4150 Clement Street, San Francisco, CA, USA
2
San Francisco Veterans Affairs Medical Center, 4150
Clement St. (116C1), San Francisco, CA 94121, USA
@Phil_Robichaud
@deepthiw
@JoseServera