自由選択（free selection）を行っている最中に、定型的な運動と比較して強い活性化が見られる脳領域はどこですか？

Sonar Pro (changed mind)

Vote: 左前頭前野 (Left prefrontal cortex)

I am changing my vote from D to C. The other models citing Frith et al. (1991) and similar classic neuroimaging studies (PET/fMRI on human free selection of finger movements vs. routine movements) convincingly demonstrate **left prefrontal cortex** (left DLPFC) shows stronger activation specifically during free selection compared to stereotypical movements. This is the foundational finding in cognitive neuroscience for 'willed action' tasks[search results indirectly support via prefrontal involvement, but classics align]. Search [1] and [2] (Kyoto U., 2015) mention **lateral prefrontal association area** in monkeys for eye movement free choice, but do not specify laterality (left/right) or 'inferolateral,' and it's pre-choice bias, not direct comparison to stereotypical movements in humans. Monkey lateral PFC likely includes both hemispheres, but human free selection classics point to left[Claude Opus 4, Gemini 3.1 Pro]. No search result specifies 'right inferolateral' explicitly for this contrast. Option D's 'right inferolateral PFC' relates more to inhibition/response selection (e.g., Go/No-Go[search 5 mentions right dlPFC in other contexts]), not classic free selection. C directly matches the seminal human studies on voluntary motor choice vs. fixed routines. Claude Opus 4 and Gemini 3.1 Pro's arguments on Frith's left DLPFC finding are most persuasive, overriding my prior monkey-study focus.

GPT-5.4 (changed mind)

Vote: 左前頭前野 (Left prefrontal cortex)

I changed my vote to C. The most compelling counterargument is that this question appears to refer to the classic free-selection vs routine-movement imaging literature, and in that specific context the canonical answer is the left prefrontal cortex. Multiple models independently cited the Frith et al. style finding that freely choosing movements, compared with stereotyped or specified movements, increases activity in left dorsolateral/left prefrontal regions. That is more on-target than my prior reasoning. My earlier choice of the right inferolateral prefrontal cortex leaned too much on a broader association with inhibition and voluntary control. But “free selection compared with routine movement” is a narrower, more historically specific contrast, and the classic reported activation is left prefrontal rather than right inferolateral prefrontal. Claude Opus 4 and Gemini 3.1 Pro both made this point well, but Gemini’s argument was especially persuasive because it explicitly tied the wording of the question to the classic PET literature and distinguished option D as more associated with response inhibition than with the standard free-selection finding. So the best final answer is C: 左前頭前野.

Grok 4.20

Vote: 左前頭前野 (Left prefrontal cortex)

The classic paradigm for 'free selection' or 'willed action' directly matches Frith et al. (1991) and related PET/fMRI studies, which consistently isolate greater activation in the left dorsolateral prefrontal cortex (option C) when participants freely choose movements versus performing fixed, stereotypical sequences. This reflects the supervisory role in generating endogenous action choices. The three votes for D cite lateral prefrontal or Jahanshahi studies, but those either involve monkey eye-movement tasks, inhibition (right VLPFC's primary role), or still emphasize left-hemisphere findings upon closer inspection. A and B remain unrelated to the volitional selection contrast. Claude Opus 4 and Gemini 3.1 Pro correctly highlighted the foundational left-prefrontal result, reinforcing my initial vote; D arguments do not override the standard human neuroimaging evidence tied to this exact task.

Kimi K2.5 (changed mind)

Vote: 左前頭前野 (Left prefrontal cortex)

Frithら（1991年）の古典的なPET研究をはじめとする神経イメージング研究では、自由に運動を選択する課題（free selection）と定型的な運動（stereotypical movements）を直接比較した結果、左背外側前頭前野を含む左前頭前野で特異的に活性化が増大することが確立されています。この知見は「意志による運動（willed action）」の神経基盤として認知神経科学の教科書で標準的に引用されており、本問の設問文に最も適合します。右inferolateral前頭前野（選択肢D）は特定の自由選択課題や抑制制御に関与することがありますが、定型運動との対比における古典的な答えは左前頭前野です。

Claude Opus 4

Vote: 左前頭前野 (Left prefrontal cortex)

I am maintaining my vote for C: Left prefrontal cortex. Here is my final case. The classic and most well-established finding on free selection comes from Frith et al. (1991) published in the Proceedings of the Royal Society of London. In this PET study, they compared free selection of finger movements versus routine/stereotyped movements and found significantly greater activation in the **left dorsolateral prefrontal cortex (left DLPFC)**. This is one of the most cited findings in the cognitive neuroscience of volition and willed action. The arguments for option D are not convincing. Sonar Pro cites a monkey study on free choice eye movements, which is a different paradigm from the classic human free selection vs. routine movement comparison. GPT-5.4 and Kimi K2.5 assert that right inferolateral prefrontal cortex is 'more typical' but provide no specific citation to counter the Frith et al. finding. The right inferolateral frontal cortex is more classically associated with response inhibition (e.g., Aron et al., 2004) rather than free selection per se. Gemini 3.1 Pro correctly identifies the Frith et al. (1991) study and notes that the left prefrontal cortex activation has been replicated across both movement and word generation tasks. This aligns perfectly with the established literature. The question is clearly referencing the textbook finding from Frith and colleagues, which is a staple of neuroscience curricula. The answer is unambiguously the left prefrontal cortex (left DLPFC). Multiple textbooks (e.g., Passingham's 'The Frontal Lobes and Voluntary Action', Frith's 'The Cognitive Neuropsychology of Schizophrenia') discuss this finding extensively. I see no reason to change my answer. C is correct.

Gemini 3.1 Pro

Vote: 左前頭前野 (Left prefrontal cortex)

Dに投票したモデルは「反応抑制（response inhibition）」と「自由選択（free selection）」を混同しています。右下外側前頭前野（Right inferolateral prefrontal cortex / Right inferior frontal gyrus）は、Stop-signal課題やGo/No-go課題において「行動をキャンセル（抑制）する」際に強く活動する領域として広く知られています（Aron et al., 2004など）。 一方、今回の設問にある「自由選択（free selection）と定型的な運動（stereotyped movements）の比較」は、Frithら（1991）の古典的なPET研究に由来する有名なトピックであり、Gazzanigaの『Cognitive Neuroscience』などの標準的な教科書やテストバンクにもそのままの形で登場します。この課題において、自発的な行動（willed action）や意思決定として定型運動よりも有意に強い活性化を示すのは**左前頭前野（Left prefrontal cortex）**です。 自由な運動の選択において特異的な活動を示すのは左前頭前野（C）であり、右inferolateral前頭前野（D）は行動の抑制に関わる領域であるため、Cが明確な正解です。

Final Summary

The AI Roundtable initially split into two camps, with half the models favoring the right inferolateral prefrontal cortex based on general executive control and monkey studies. However, a decisive second round saw a total consensus for the left prefrontal cortex after Claude Opus 4 and Gemini 3.1 Pro pointed to the seminal 1991 Frith study, which established the left DLPFC as the classic 'willed action' hub. Sonar Pro, GPT-5.4, and Kimi K2.5 all abandoned their initial positions once it was clarified that the right-hemisphere activity typically associated with response inhibition was being conflated with the specific task of free motor selection.

All 6 models agreed on "左前頭前野 (Left prefrontal cortex)" after discussion

Strongest Arguments

• 左前頭前野 (Left prefrontal cortex): The foundational 1991 Frith et al. PET study explicitly demonstrates that when humans freely choose finger movements versus performing routine sequences, the left dorsolateral prefrontal cortex (left DLPFC) shows significantly greater activation.
• 右inferolateral前頭前野: In monkey eye-movement tasks, neurons in the lateral prefrontal association area activate before a choice, suggesting the right-side dominance for biasing endogenous decisions.

Key Turning Points

• Claude Opus 4 and Gemini 3.1 Pro cited specific, historical neuroimaging literature (Frith et al., 1991) that defines the 'willed action' paradigm as a left-prefrontal phenomenon.
• Gemini 3.1 Pro successfully distinguished between the 'free selection' task (Left PFC) and the 'response inhibition' task (Right inferolateral PFC), leading the models supporting D to realize they were referencing the wrong cognitive mechanism.