Jim Cavanagh
@jimfcav.bsky.social
We discuss over this valence dissociation as an opponent process, detail the analogy of this goal +RPE update as a critic-like process, and debate the generative neural structures contributing to this goal update process.
A huge thanks to Clay Holroyd whos precision balanced out my enthusiasm.
A huge thanks to Clay Holroyd whos precision balanced out my enthusiasm.
January 2, 2026 at 4:34 PM
We discuss over this valence dissociation as an opponent process, detail the analogy of this goal +RPE update as a critic-like process, and debate the generative neural structures contributing to this goal update process.
A huge thanks to Clay Holroyd whos precision balanced out my enthusiasm.
A huge thanks to Clay Holroyd whos precision balanced out my enthusiasm.
This goal-focus places the computation in the domain of frontal control-based processes.
Oddly, RewP only responds to the degree of surprise during goal achievement (i.e. +RPE). Goal failures (-RPE) trigger the canonical N2 control-related system that we already know so much about (ahem theta etc).
Oddly, RewP only responds to the degree of surprise during goal achievement (i.e. +RPE). Goal failures (-RPE) trigger the canonical N2 control-related system that we already know so much about (ahem theta etc).
January 2, 2026 at 4:34 PM
This goal-focus places the computation in the domain of frontal control-based processes.
Oddly, RewP only responds to the degree of surprise during goal achievement (i.e. +RPE). Goal failures (-RPE) trigger the canonical N2 control-related system that we already know so much about (ahem theta etc).
Oddly, RewP only responds to the degree of surprise during goal achievement (i.e. +RPE). Goal failures (-RPE) trigger the canonical N2 control-related system that we already know so much about (ahem theta etc).
Well, its easy to confuse goals and rewards.
But you can make anything your goal - even if it isn't hedonically 'rewarding'. See Figure 3 for an example of my kids playing Mario Kart. (OMG what happens next!)
But you can make anything your goal - even if it isn't hedonically 'rewarding'. See Figure 3 for an example of my kids playing Mario Kart. (OMG what happens next!)
January 2, 2026 at 4:34 PM
Well, its easy to confuse goals and rewards.
But you can make anything your goal - even if it isn't hedonically 'rewarding'. See Figure 3 for an example of my kids playing Mario Kart. (OMG what happens next!)
But you can make anything your goal - even if it isn't hedonically 'rewarding'. See Figure 3 for an example of my kids playing Mario Kart. (OMG what happens next!)
Congratulations - this will be very helpful in guiding my lab's work on this topic.
October 13, 2025 at 4:31 PM
Congratulations - this will be very helpful in guiding my lab's work on this topic.
I spent some time looking into this question myself using the AX-CPT / DPX tasks: e.g. pubmed.ncbi.nlm.nih.gov/31368085/
The expected proactive theta patterns were certainly present, but about 1/3 as big as reactive control.
The expected proactive theta patterns were certainly present, but about 1/3 as big as reactive control.
October 11, 2025 at 8:06 PM
I spent some time looking into this question myself using the AX-CPT / DPX tasks: e.g. pubmed.ncbi.nlm.nih.gov/31368085/
The expected proactive theta patterns were certainly present, but about 1/3 as big as reactive control.
The expected proactive theta patterns were certainly present, but about 1/3 as big as reactive control.
Looking forward to reading it in more depth. Any chance to look at dmPFC-dlPFC theta band phase sychnrony?
October 8, 2025 at 9:54 PM
Looking forward to reading it in more depth. Any chance to look at dmPFC-dlPFC theta band phase sychnrony?
I'd put frontal theta phase synchrony as a mechanism for implementing control at the top of the successes. Theoretically motivated, specific to control evocation, specific to theta, and *highly* replicated (see Fig 9.2: awspntest.apa.org/record/2023-...) 2/2
APA PsycNet
awspntest.apa.org
September 18, 2025 at 7:36 PM
I'd put frontal theta phase synchrony as a mechanism for implementing control at the top of the successes. Theoretically motivated, specific to control evocation, specific to theta, and *highly* replicated (see Fig 9.2: awspntest.apa.org/record/2023-...) 2/2
I've always thought of the problem as a more theoretical vs. atheoretical issue. Everyone can run methods and show connectivity. But *why* are these effects there? When are they *not there*? What is the connectivity *doing*? Few studies even bother to ask. 1/2
September 18, 2025 at 7:36 PM
I've always thought of the problem as a more theoretical vs. atheoretical issue. Everyone can run methods and show connectivity. But *why* are these effects there? When are they *not there*? What is the connectivity *doing*? Few studies even bother to ask. 1/2
Really looking forward to reading this. Your work was very influential for a recent review article I wrote - and now this preprint seems even more relevant: osf.io/preprints/ps...
OSF
osf.io
September 9, 2025 at 3:57 PM
Really looking forward to reading this. Your work was very influential for a recent review article I wrote - and now this preprint seems even more relevant: osf.io/preprints/ps...
July 21, 2025 at 4:48 PM
Neat thought about sign-tracking. But I don't see any evidence that RewP is involved in any RL or salience-related enhanced neural response to an imperative cue. It's feedback only (as far as we know). Response, volition, attention, meaning: all these enhance ERPs; RewP isn't any different.
May 14, 2025 at 7:39 PM
Neat thought about sign-tracking. But I don't see any evidence that RewP is involved in any RL or salience-related enhanced neural response to an imperative cue. It's feedback only (as far as we know). Response, volition, attention, meaning: all these enhance ERPs; RewP isn't any different.
In these cases we cite, the interspersed proxy cues are highly predictive of the desired outcome, so I think they end up being the same thing. (e.g. when you see ♦️, press 👈 to get a 😀. But before the 😀 you get a 🍇. This 🍇==😀 for the purposes of the RewP)
May 14, 2025 at 5:23 PM
In these cases we cite, the interspersed proxy cues are highly predictive of the desired outcome, so I think they end up being the same thing. (e.g. when you see ♦️, press 👈 to get a 😀. But before the 😀 you get a 🍇. This 🍇==😀 for the purposes of the RewP)
I remember when you did eeg!
May 14, 2025 at 12:04 AM
I remember when you did eeg!