The Neurocritic: Machine Yearning - Sad Robots and Prolonged Grief

Crying Robot, by Mr.A

What is 'machine yearning'?

Intense longing exhibited by cartoon robots?

Or a clever pun that describes a network analysis of prolonged grief symptoms? (Malgaroli et al., 2022).

My late wife was a writer who was very fond of robots and Futurama. This post is an opportunity to incorporate them all into a brief narrative about the computational psychiatry of prolonged grief disorder.

Concert goers booed the avant garde tuba playing robot, but it wasn't programmed to feel sad, or to stop. #Citybots #Warsaw #TwitterFiction
— S. Kay (@blueberrio) March 16, 2014

Prolonged Grief Disorder (PGD) is an ICD-11 diagnosis that overlaps with other formulations of “pathological grief” including “persistent complex bereavement disorder” (PCBD) and “complicated grief”. The ICD-11 definition and symptoms can be found here.

“Prolonged grief disorder is a disturbance in which, following the death of a partner, parent, child, or other person close to the bereaved, there is persistent and pervasive grief response characterised by longing for the deceased or persistent preoccupation with the deceased accompanied by intense emotional pain (e.g. sadness, guilt, anger, denial, blame, difficulty accepting the death, feeling one has lost a part of one’s self, an inability to experience positive mood, emotional numbness, difficulty in engaging with social or other activities).”

The symptoms must persist for at least six months, with consideration of differing cultural norms for bereavement. The overlap between PGD and PCBD is shown in purple in the figure below.

– click on image for a larger view –

Infographic. Disturbed grief: prolonged grief disorder and persistent complex bereavement disorder (BMJ).

Painful and prolonged yearning for the lost loved one is an obvious core symptom.

She pines for her partner and cheese enchiladas in mole sauce at that festive restaurant, celebrating life and Cinco de Mayo. @blueberrio
— Seven By Twenty (archive only) (@7x20) May 5, 2011

[NOTE: A celebration of life is viewed as a more adaptive response.]

But there's so much more... An aptly titled paper by Lenferink and Eisma (2018) indicated there are 37,650 ways to have “persistent complex bereavement disorder” yet only 48 ways to have “prolonged grief disorder”. ¹ For a PCBD diagnosis (a DSM-5 construct), an individual must have at least one of four Cluster I symptoms and at least six of twelve Cluster II symptoms. For a a PGD diagnosis, an individual must have at least one of two Cluster I symptoms and at least three of five Cluster II symptoms. This is a rather stunning degree of symptom heterogeneity, and much more complicated than the Infographic suggests. It might even limit the usefulness of these diagnostic categories, which may be true of psychiatric diagnoses in general (see RDoC).

Here is where computational methods may assist with understanding symptom profiles and trajectories of grief (Malgaroli et al., 2022). Data driven analyses using unsupervised machine learning methods can identify patterns within the heterogeneity. Network analyses have identified the centrality of loneliness (Fried et al., 2015) and a social/identity symptom cluster that includes role confusion and meaninglessness (Malgaroli et al., 2018). In the review article, the authors thought this was novel and stated that, “A more surprising and emergent result showed meaninglessness and role confusion to be strongly central elements.”

But I am not surprised at all.

My wife (@blueberrio) wrote microfiction — miniature standalone stories as well as longer collections. Her book Reliant was an entire apocalypse in tweets. After the nuclear crisis occurs in Part 2, the main character channels the agony of loss in a heartbreaking (and prophetic) way.